Amazon will no longer build four new data centers on the site of the George Washington University's campus in Loudoun County, just outside of D.C., and will instead repurpose the majority of the 122-acre space for mixed-use and residential development following significant pushback from the local community. It will also donate 25 acres of the property to Loudoun County to be used for a community arts center and new recreational facility. On top of all that, it'll even give $5 million to the county's fund for housing, reports The Washington Post.

That's quite the turnaround compared to plans to build in a county that is already dubbed "Data Center Alley" because it has the largest concentration of data centers on the planet. But residents' concerns about the noise generated by such data centers and the incredible amount of power that they require to function have seen the county's approach to new data center buildouts change of late. Amazon first submitted its proposal to build the new facilities in July, and Loudoun County Board of Supervisors Chair Phyllis J. Randall and Amazon met in September, where it was put to the company that it may be better to find an alternative use for the land.

Josh Weissman, Amazon’s vice president in charge of data center development, said the company had jumped at the chance to purchase the former university campus when the opportunity arose, but added that the company considered the county’s “pointed concerns about data center development at this particular site.” Weissman added that “While we see the property’s legal development potential differently than the County does, we concluded that the best outcome for all is to come together on a shared vision for this campus rather than engage in costly litigation.”

The building of new data centers has always been a necessity given our growing use of cloud-based services and storage systems. But the recent boom in AI use has seen demand for additional, and larger, facilities ratcheted up to new levels. But with that demand comes opposition, with local communities wary of the impact these mammoth facilities can have on the areas in which they live, including spikes in the price of electricity. Officials have begun taking steps to try to grease the wheels of data center construction. Just days before this news broke, President Donald Trump called for data center builders and AI hyperscalers to pay local communities a dividend amid energy-related price increases. In a social media post, Trump went on to say that data centers would be required not only to make their own electricity but also to give or sell excess energy back to the local community. Water shortages and deforestation have also become key issues affecting the communities in which data centers are being placed.

Back in Loudoun County, officials pointed to what could be achieved when choosing dialog over confrontation. “By choosing dialogue over reflexive disagreement, Loudoun County and Amazon created an opportunity for the County,” Randall reportedly said via statement. “This agreement provides new places for our Loudoun residents to gather, create, and play. And, most importantly to me, we will provide more attainable homes for working families.”

Export controls have kept ASML's most advanced EUV lithography machines out of China. Instead, the country's chipmaking industry has relied on less-capable deep ultraviolet (DUV) lithography machines. By using multiple exposures on critical chip layers, these older machines are still capable of producing performant silicon. However, as multiple patterns are applied there are more chances for defects, longer manufacturing times, and higher bills. We spoke to experts like Chris Mack, co-founder and CTO of semiconductor firm Fractilia, to find out how much further DUV can go.

For years, extreme ultraviolet (EUV) lithography has been seen as China’s biggest choke point in the country’s attempts to build cutting-edge chips. The machines, made exclusively by Dutch manufacturer ASML, use 13.5nm-wavelength light to print some of the most intricate features found on modern processors. Sales of ASML’s EUV systems are subject to export restrictions, forcing Chinese manufacturers to squeeze ever more out of older deep ultraviolet, or DUV, lithography instead. The EUV gap alone between China and the West is at least half a decade, according to some analysts.

But even without access to EUV machines, China has gotten surprisingly far in recreating what it offers through existing technology.

Teardowns of Huawei’s Mate 60 Pro in 2023 found the phone’s Kirin 9000S processor had been made by SMIC using a second-generation 7nm-class manufacturing process without EUV. More recently, analysis of Huawei’s Kirin 9030 found SMIC had pushed its DUV-based process closer still to the density associated with 5nm manufacturing – although it continued to lag behind the leading processes used by TSMC and Samsung.

Multiple patterning

The trick is multiple patterning. Rather than trying to print an overly intricate chip layer in a single pass, Chinese manufacturers are dividing up the complex design into several simpler patterns. Repeat the process enough times and you can etch up something approximating the proper layer you intended.

The trick works. But according to lithography expert Chris Mack, co-founder and CTO of Fractilia, asking whether a chip can be made without EUV overlooks the trade-offs manufacturers make in order to do it.

“When this transition first began, what everyone in the industry said as conventional wisdom is that cost-wise, three exposures on a 193 immersion tool cost about the same as one exposure on an EUV tool,” Mack said in an interview with Tom’s Hardware Premium. “At three, it was about even,” Mack explained. “But as soon as you hit four masks – 193 masks being replaced by one EUV mask – then EUV became more cost-effective.”

It’s also likely to produce a higher quality product with fewer defects. Each time a wafer is patterned, that new pattern has to line up perfectly with everything that’s already been etched beneath it. The more individual exposures needed to produce a single layer, the more opportunities there are for things to go wrong. “If you have, say, four 193 masks, you've got all those different overlays, and you've got all these different process variations that start to accumulate,” Mack said.

Chip designers can try and avoid that by leaving more space around features so the finished processor can actually be manufactured reliably. But that means the chip ends up bigger – and bigger chips become expensive very quickly. Making each die larger means fewer fit on every wafer, and wafers cost a lot of money – tens of thousands of dollars each.

“You don't sell wafers, you sell die,” Mack said. “The question is: how many die can I fit onto that wafer, and how many of those die yield, and therefore how many chips can I sell per wafer?”

The efficiency gap

There is some evidence of what that efficiency gap can look like in practice – albeit in memory rather than logic chips. In a June report, Morgan Stanley estimated that Chinese DRAM manufacturer CXMT was producing only around one-third as many bits per wafer as Western competitors. The analysts believed the older process node only accounted for around a third of that shortfall, but said access to lithography was a “hard ceiling” on Chinese memory development. “For advanced logic chips, in theory there could be other innovations to engineer around EUV, but they are not likely to be very efficient and effective,” said Douglas Fuller, a researcher at Copenhagen Business School, in comments to Tom’s Hardware Premium.

It is also why claims that China has reached a particular semiconductor process “node” need to be treated with caution. “If somebody says it's X node – the seven nanometer node, the five nanometer node, the three nanometer node – these words almost have no meaning,” Mack said.

That can be seen in SMIC’s latest chips. Reporting suggests its N+3 process represents meaningful progress towards 5nm-class manufacturing – but the actual measurements show it remains significantly less scaled than established 5nm processes from Samsung and TSMC. That doesn’t mean DUV has reached the end of its useful exploitation.

Stretching DUV's efficacy

Micron is an example of how far DUV can be stretched, producing competitive DRAM chips using DUV, multiple patterning and tight process control. Its 1-alpha generation of chips was developed without EUV. “They pushed 193 farther than any of the other memory makers did,” Mack said. “They were able to make very competitive chips with 100% 193 compared to Samsung and Hynix making them with many layers of EUV.”

Morgan Stanley believes CXMT’s planned move to 1-alpha DRAM – which it wants to use for HBM3e – is achievable without access to EUV, citing Micron as evidence. Indeed, “by the end of the decade, advanced packaging will allow firms to use DUV to produce the latest generations of DRAM,” reckons Fuller.

Micron also designed its products around the manufacturing technology available to it, rather than simply trying to recreate an EUV process on older equipment – something which could be a useful lesson for Chinese manufacturers. “If the entire ecosystem of 193 immersion were available to China,” Mack explained, “then they could probably push a long way with 193 multiple patterning, especially if they optimise designs for the fact that they're going to be using that.”

But Micron was making memory, whose highly repetitive structures make close co-ordination between chip design and manufacturing easier than it is for the much more complicated logic processors found in phones and computers.

“Memory is a lot easier than logic,” Mack said.

And even Micron eventually concluded that EUV was worth using. Its newer 1-gamma DRAM process uses EUV lithography, with Micron saying the technology helps it achieve more than 30% greater bit density per wafer than its preceding 1-beta generation.

EUV is hardly a magic wand, either. Mack points to stochastic effects – tiny random variations caused by the small number of photons involved in an exposure – which mean manufacturers can improve pattern control by increasing the EUV exposure dose. But it also pushes the cost back up quickly.

Because of that, Mack believes there is ultimately a limit to how far multiple patterning can economically substitute for EUV. “I would be hard pressed to really believe that a three nanometer or a two nanometer node is going to be economically viable without EUV,” he said.

That doesn’t mean Chinese chipmakers simply hit a technological wall and stop. They can keep pushing DUV while accepting larger dies, poorer yields, longer manufacturing times – or some combination of all three.

They can also throw capacity and money at part of the problem. Morgan Stanley assumes CXMT remains stuck on its 1-alpha process without EUV, yet still increases its share of global DRAM bit supply from 14% in 2026 to 16% in 2028 simply by building more greenfield capacity. The bank also expects China’s AI chip self-sufficiency to rise from 42% in 2025 to 70% by 2030 through new advanced-node capacity, improving yields and heavy state support – while EUV, HBM and advanced software leave a persistent gap. And so without EUV, the experts reckon China can only get partway to bridging the gap with the West. ““I would be hard pressed,” said Mack, “to really believe that a three nanometer or a two nanometer node is going to be economically viable without EUV.”

Lava, an Israeli security firm, says that it has been able to identify 32 Nvidia Blackwell Ultra B300 chips within the network of China Telecom, a state-owned company that provides telecommunications services within the country. These same chips have already been barred from export by U.S. export controls, a report by The Information notes, which begs the question of how they have turned up in a Chinese data center. Such chips can only be sold to whitelisted buyers who agree not to export them to sanctioned parties, including those in China.

However, banning companies from buying Nvidia's high-performance GPUs and AI data center chips hasn't always worked. In 2025, it was reported that Chinese companies smuggled $1 billion worth of Nvidia AI chips into the country in three months. At the time, the B200 chips were the most sought-after, but some distributors were already gearing up to sell B300-series chips when they became available.

While the latest finding shows that the proliferation of export-controlled chips into China continues to be an issue, Nvidia has downplayed the report but didn't deny the findings. "Governments, researchers, and media have all found insignificant diversion of controlled US products, less than one half of one percent in recent court cases — a drop in the bucket compared to the ocean of domestic compute China already has," an Nvidia spokesperson told Tom's Hardware. "With tens of millions of GPUs in service in approved countries, this report again demonstrates that smuggling is a nonfactor, blocked by governments on both sides of the globe."

It's unknown how B300 chips found their way into China Telecom's server racks, nor can Lava confirm exactly where within the company's network they are located. But they are there, with Lava identifying them via Nvidia's own DCGM Exporter tool. The software can be installed by data center operators and is used to monitor server health and performance, and while designed for internal use, it can oftentimes be accessed beyond the data center's network. Oddly, it doesn't always require password authentication, either. Nvidia previously confirmed it was working on a tool that would allow its chips to be located, but it is yet to be released.

It's easy to see why Chinese firms are keen to get their hands on the Blackwell Ultra B300. Announced in March of last year, the data center GPU is one of the most powerful on the planet, offering huge performance gains with 50% more memory and FP4 compute than the previous-generation B200. At the time, Nvidia touted that the chip was "built for the age of reasoning," while name-checking more sophisticated AI LLMs like DeepSeek R1. At the time of its announcement, the B300 was expected to ship by the end of 2025, and it didn't take long for it to become hot property. In August 2026, nine people were indicted in Taiwan over the illegal export of B300 GPUs to China. This came just days after Supermicro confirmed that it had fired several employees following an investigation into smuggling. The complicated smuggling process included the use of a Taiwanese server trading firm called Flying Tiget Tech — a firm with whitelisted status for GPU purchases.

Nvidia's chips have proven so vital to AI data centers that demand for even its older silicon has caused skyrocketing prices. The company's A100 servers have tripled in price, despite being five years old, and can cost up to $82.,000. The ongoing crackdown on Nvidia chip smuggling has even seen some resort to retooling gaming GPUs to run AI workloads where possible.

The FCC recently imposed a new rule allowing it to ban devices based on the internal chips instead of basing bans on individual products or brands, thus closing the loophole of rebranded products with parts from Chinese manufacturers on US export restriction lists. Using this new rule, the FCC has banned two devices from being sold and marketed in the US for the first time: a dashboard camera and a wireless video transmitter. As reported by The Verge, these devices were banned due to the HiSilicon Smart Vision SoC (System on a Chip) they contain, which features built-in detection algorithms for faces. HiSilicon is a subsidiary of Huawei, and the firm and its subsidiaries are on the FCC’s Covered List of entities that pose a risk to US national security.

The Hi3516CV610 UHD Smart Vision is an SoC for surveillance and camera products. The SoC is meant to be used for things like vehicle dashcams and wireless CCTV systems; fully equipped with WiFi, Ethernet, video encoding chips, an Arm Cortex-A7, and an NPU (Neural Processing Unit) for AI features. The NPU is what the FCC is concerned about, explaining that it is a device that contains logic-bearing “smart” components produced by entities on the Covered List.

On the SoC’s data sheet, the Intelligent Video Engine (IVE) uses the NPU that runs a built-in dedicated multimodal large model and transformer feature acceleration. The transformer feature acceleration allows the AI model to understand the relationship between objects from one video frame to another, keeping “attention” between related objects: pets, faces, human figures, and parcels, as the data sheet advertises. For consumers, this potentially means the AI in a surveillance camera can understand whether it is your pet that is escaping, or an unknown person is stealing your parcel after understanding that it was placed by a courier worker.

However, the FCC believes the logic-bearing AI in the SoC can do something much more nefarious, surveying US citizens and logging faces for foreign surveillance. Huawei already had a previous AI chip scandal, using shell companies to deceive TSMC into producing AI chips for them, going against the US’ request for TSMC to halt supplies to a list of Chinese customers; among other Huawei scandals leading to other product bans in the US.

Deepseek just upgraded its V4.1 Flash open-weight model to revision 1017. What looks like a mere point revision may actually be a significant step forward in model efficiency. On the product page, Deepseek uses the slogan "wish-driven programming made real," a likely reference to the colloquial "vibe coding," implying the model is a good fit for open-ended tasks with broad descriptions. However, while the model is indeed cheap and scores well, it can be hallucination-prone.

LLMs are roughly divided into two categories: closed commercial efforts from frontier labs, whose models are expensive to run but deliver excellent results, and Chinese open-weight competitors that broadly exchange reliability for a much cheaper running cost, particularly for long-running agentic tasks.

Deepseek V4.1 Flash (DSF) is a Mixture-of-Experts (MoE) model with 552B total parameters, 16B of which are active during runs. The model is multimodal, accepts both text and images as inputs, and its context window length measures the now-standard 1 million tokens.

According to Deepseek's internal benchmarks, the model slots in neatly between GPT-5.6 Sol and Claude Opus 5, even beating Anthropic's LLM in task tests. That's already a strong showing, but this latest revision of DSF shines in the performance-per-task base and token efficiency. The upgrade from V4 Flash to V4.1 Flash significantly slashed hardware requirements, with V4.1 using only one-quarter of the HBM and one-eighth of SSD storage for its KV cache compared to its predecessor, spending only 890 bytes per token entry.

The company illustrates the point with a cost-per-task graph based on DeepSWE's dataset from September 2, adding its own data points for both DSF and DSF-1017. Both variants cluster in the desirable top-left area, and even low-effort runs of DSF-1017 offer a task completion rate in the ballpark of "max" settings for pricey competitors. It's worth noting that DeepSWE is only one of many benchmarks, but this is nonetheless an impressive result.

Even at Deepseek's own pricing (remember, this is an open-weight model you can run anywhere), the price at off-peak times per million tokens is just $0.15 for input and $0.6 for output. On cache hits, the input token cost drops to a mere $0.003. It's worth keeping in mind that third-party providers have undercut Deepseek API in the past, so those figures could go lower if you shop around inference providers. Deepseek also offers search functionality at $0.75 per 1,000 queries, an amount that undercuts competitors, with the closest one being Parallel at $1.

DeepSWE isn't the only benchmark around, either. Artificial Analysis ranks DSF-1017 (max) as only #2 on its intelligence index within the large open-weight class, with a score of 45. For reference, the most seemingly intelligent model, Claude Opus 5.5 (max with fallback), rings in at 58. It's one of the fastest models in its category, too, and its cost-per-task is edged out only by Xiaomi's MiMo V2.6-Pro.

This all sounds too good to be true, but everyone knows the adage of "fast, cheap, and reliable, pick two." So what's the catch? In one word: hallucination. Just because a model is smart doesn't mean it's reliable. Broadly speaking, users prefer models that don't answer when they lack supporting facts, and DSF-1017 doesn't seem very good at knowing what it doesn't know.

Artificial Analysis' Omniscience benchmark is quite clear on this matter, ranking the model among the lowest. It answered questions correctly 43% of the time, which is among the best for open-weight models, but inferior to the proprietary competition. That would be perfectly acceptable if it skipped answers without a foundation, but it hallucinated 67% of non-answers.

The benchmark's main index grants points for good answers and deducts them for wrong ones, and non-answers don't count. DSF-1017 gets 5 points, compared to Kimi K3-max's 20, with the frontier models coming in at 42+. Ultimately, this means DFS-1017 is by no means incompetent — much the opposite —but one should be careful about which tasks to assign to the model.

Tasks with verifiable outputs like programming, OCR with review, and format conversions will be fine, as the agent can run verification steps against its work and try again. However, open-ended research, looking up facts, and interpreting or drafting highly precise documents are tasks better suited to a frontier model — or, you know, a human — as hallucinations likely won't be properly reined in and results won't be very trustworthy.

"Fast and loose" seems to be a forming trend for highly competent open-weight models, too. Several months ago, when Kimi K3 was released, everyone raved about how effective and cheap it was, and how it proposed to offer frontier-level intelligence for a fraction of the price. My personal experience with Kimi was short-lived, as even short but precise fact look-ups came up with hallucinated information, and Artificial Analysis' data seems to back this observation up.

Nevertheless, open-weight models remain important both to make AI agents more accessible in absolute dollar terms and potentially as a tool to drive down the cost of intelligence when facing down potential price rises from frontier labs.

A new report claims that major U.S. banks have now detected AI agents accessing customer accounts, and that some are even attempting to make transactions. Axios reports that it’s hard to differentiate between an actual human and an unauthorized bot, which is why some institutions, including Morgan Stanley, Bank of America, JPMorgan, and Wells Fargo, are blocking the technology and changing their terms of service to account for this.

While an AI agent or AI assistant can access an account with a user’s permission and through the use of their login credentials, this often goes against bank policy that requires users to maintain control of their account and not share it with anyone else. One of the fears that executives have is when an AI agent with login access takes unauthorized steps, like spending money or moving funds around without the explicit instructions of the account owner. There have even been situations where an AI agent called the bank directly and took actions independent of direct user command. Even though it wasn’t explicitly told to do that transaction, it considered it a crucial step to accomplish the task assigned to it.

One hypothetical example of this is when a user tells an AI agent to purchase an item but gives it a credit card with no available limit. The AI agent could then have accessed the user’s bank account to move some funds from whatever account it sees that has available funds to then effect the transaction, even if the user did not explicitly tell them that they could use the money in that said account. This is just one example where an AI agent’s transaction could go wrong, which is why several U.S. banks are now restricting third-party AI agents from accessing their platforms. Nevertheless, they still recognize that their customers might want to use tools like AI agents to monitor their finances. This is why some institutions have given or are moving towards read-only access for AI agents, which is something that they’ve been doing for years already.

The move comes against a background of burgeoning AI agent use, coupled with the risk that agents can and often do act beyond their remit, making mistakes or performing tasks without instruction, often with disastrous results. An unsupervised AI agent recently deleted a TV show's entire edit weeks before broadcast, forcing the program off air. AI agents are notorious for deleting databases and directories, making it a small leap to imagine they cause some havoc with unfettered access to a bank account.

Financial institutions have deep experience when it comes to employing new technology while ensuring that their critical infrastructure remains safe — something that AI labs lack, according to a former OpenAI safety employee. Despite that, banks are still concerned about the speed at which AI technology is developing, fearing that the technology’s capabilities might outrun their ability to control it, which is why they’re preparing for worst-case scenarios and even planning for “day after” scenarios after a catastrophic AI event.

A Hacker News user who shared news about the recent PHP vulnerabilities revealed by ChatGPT Sol before an embargo, which they claimed to be unaware of, has taken to the platform to apologize for the post. According to petecooper, they shared the CVEs because they wanted to alert others who follow the four-week PHP release cycle about the importance and urgency of the matter, but they weren’t aware of the news blackout that went with the two identified vulnerabilities.

The user said they have "No connection or affiliation to any of the organisations involved." Noting, "I am a self-employed consultant for my day job, and a solo founder of a mental health non-profit on the side." Petecooper's original post about the vulnerability was among the first to break the NDA, though they claim they were unaware.

“I posted here about the then-upcoming PHP RCEs related to CVE-2026-109618 & CVE-2026-109619. I appreciate this has had considerable knock-on effects with the schedule for co-ordinated disclosure, and created additional workload. I am truly sorry. I take full responsibility for my actions,” petecooper wrote on the platform. “It was not remotely my intention to cause any of the subsequent side effects. At the time of posting, I didn’t know about any embargo. My contact had told me there was no embargo, and her contact had told her the same thing.”

CVE-2026-109618 and CVE-2026-109619 are two critical vulnerabilities, both scored at 9.2. It seemed that the PHP Foundation planned for a controlled public release, allowing affected websites and applications to be patched before the vulnerabilities became widely known. Unfortunately, it seemed that some recipients of the warning failed to share the embargo that came alongside the CVEs, spreading the news earlier than expected.

It should be noted that petecooper apparently wasn’t the only one who leaked the news early. Multiple leaks had already been circulating on social media and other platforms, which forced the PHP Foundation to move its announcement up by one day. Despite that, the organization mentioned that Chainguard broke the embargo by more than six hours. Dan Lorenc, the co-founder and CEO of Chainguard, defended the move on LinkedIn, saying that he made the call after talking to several other researchers and immediately notified the group after releasing the fix publicly, and not just his customers.

The vulnerabilities tracked as CVE-2026-109618 and CVE-2026-109619 were bugs in the PHP runtime itself that were discovered using AI tools, not attacks by rogue AI. Still, the abilities of this technology could make it easier for both security researchers and bad actors to find vulnerabilities that could have been lurking in widely used code for years, if not decades. This makes it a race to find and deploy a fix before they can be exploited.

CyberCartel, a hacking group with a long history of targeting Latin American businesses, has been using AI agents, including SpaceX's Grok and Anthropic's Claude, to automate attacks on financial institutions such as banks and mobile payment apps. Passwords and customer data have been stolen, while individuals' bank accounts have also been accessed with little input from criminals in the latest example of criminals turning to AI-powered cyberattacks. That research comes from a cybersecurity startup called Straiker, with its findings shared by The Information.

While CyberCartel has been an active threat since at least 2012, its new leveraging of the power of AI has leveled up its attacks in recent months. Now, the research shows that the group can attack more businesses at a speed previously not possible when using human-powered hacks. Straiker shared its research with both SpaceX and Anthropic, and both firms are working to lock CyberCartel out of their services. The latter has already been able to block access to Claude. Chinese lab Moonshot AI's agent has also been used by the hacking group, but Straiker is still trying to get in touch with the company.

While it's true that AI models normally have protections in place to prevent them from being used to carry out cyberattacks, they've proven surprisingly easy to bypass. In this case, CyberCartel reportedly simply told the models that they were being used in test environments. If one model refused to proceed, the hackers used a different model. If that also refused, the AI agents would be tasked with deploying malware instead.

Notably, CyberCartel doesn't appear to be using Anthropic's Mythos for its recent attacks. The company caused a stir earlier this year when it announced the new model, warning that hackers could use it and models like it for attacks like those carried out by CyberCartel. Its release also saw security teams around the globe rush to address thousands of vulnerabilities that the model had identified.

This is far from the first time that Anthropic has been faced with the news that its AI is being used by bad actors. The company shared details of its AI being used to carry out a cyberattack in November 2025. Later, Claude was used to break the security of OpenAI, an Anthropic competitor in the AI space. That hack was carried out by researchers rather than criminals, at least.

Raiker's research on AI-powered attacks on Latin American financial institutions comes just days after news broke that South Korean authorities are investigating a similar issue affecting local banks. Hana Bank, KB Kookmin Bank, and Shinhan Bank are among the South Korean banks affected, with at least 25,000 Shinhan Bank customers having their personal credit information leaked.

AI companies are well aware of the risks their products pose. Nvidia, the graphics card company that now focuses on producing hardware for AI data centers, recently launched the Open Agent Safety Platform. The software platform and reference system design were developed to secure rogue AI agents via a new hardware and software security stack. It's hoped that it will go some way toward preventing agents from escaping their sandboxes, executing unauthorized code, gaining unauthorized access to critical infrastructure, or bypassing guardrails.

China-state-aligned actor Storm-0940 used ARTEX, an open-source agentic AI penetration testing framework, in its large-scale ConnectWise supply chain attack, Microsoft wrote in its Security blog. After unauthorized access to the multi-tenant application, the group attempted sign-ins at more than 2,250 Microsoft 365 tenants, with stolen ConnectWise app credentials that authenticated as the apps with no interactive user sign-in. Some of the group’s attacks reached privileged access less than three hours after initial targeting.

Automated password guessing, or password spraying, also targeted 327 accounts across eight organizations with more than 151,000 failed logons and no successes. The use of ARTEX, “a system of LLM-driven agents that run an autonomous penetration test,” sped up scanning and reconnaissance, password spraying, and OAuth application abuse. The agents can work around the clock “with little needed human interaction,” at a pace that “leaves defenders with little time to detect and respond.” It’s uncertain how the ConnectWise credentials were stolen or how many tenants were actually compromised.

Storm-0940, active since 2021, is a “China-state aligned actor” with “no known industry aliases,” and its name follows the convention for a developing actor. Industries targeted this time include aerospace and defense, semiconductor, government, and think tanks, with 14 listed in total. The activity “did not result from a compromise of or a vulnerability” in Microsoft’s systems, products, or services.

At its peak on Oct. 11, the password spray engaged in 2,900 attempts, all failures, per minute, at more than 111,000 for that day, with 48 accounts logging more than 1,000 failures each. Almost 98% of this activity was locked out. However, with control of the identities of multiple ConnectWise cloud apps, the attacking AI agents signed in to customer tenants from the same IP range using those apps’ credentials, with more than 200 sign-ins in under four minutes using four application credentials. Operations taken within some tenants include Graph and directory operations, reconnaissance, app credential changes, plus changes to identities, groups, and privileged roles.

In one tenant on Oct. 15, Huntress, a security firm, saw an app add itself to the top administrator role, the behavior behind the Microsoft Defender alert “Application self-assigned Global Administrator role.” Actions included planting new secrets set to expire in 2028. It confirmed unauthorized sign-ins for 162 tenants by Oct. 16 and, a day later, almost 1,000 Entra tenants with sign-in activity from the attacking IP, with more than 300 succeeding. The attack’s sweep through tenants in sorted order, a few seconds apart, indicated “a script working from a target list.”

What differs from many traditional attacks is the use of the open-source pentest framework that, Microsoft says, can use any AI model. Threat actors in multiple regions are using the technology, and public reporting this month tied it to breaches at South Korean financial institutions. CrowdStrike said the suspect in those separate attacks was likely based in China and allegedly used ARTEX with Anthropic’s Claude Code. ARTEX’s developer, who goes by the handle “Autumn-27,” said on GitHub that the ARTEX project “will no longer be updated” and would become closed source, Reuters reported.

ConnectWise noted the unauthorized access and suspended Cloud Backup (formerly SkyKick) as well as SaaS Security and Cloud Management as a precaution. It has started a phased restoration with recovery expected to take several days and has no indicators to date that partner backup data was compromised. Microsoft notes that an affected app, or even a successful sign-in, doesn’t necessarily mean data was accessed or taken, and advises customers to keep affected app connections disabled until formal restoration instructions are issued, and to review activity from this month where logs allow.

Google’s Finland country manager has apologized for unauthorized forest clearings near Muhos, which have turned 816 acres of forest into an expanse of mud. The American search and AI giant has also been told to cease work where it has cleared roughly 500 acres of forest, 55 miles southeast of the village of Otanmäki. Now The Guardian reports that wolves, who used to live in the forests, are roaming into villages neighboring these man-made muddy wastelands.

Local anger continues to simmer over the damage done to Finnish forests, as Google seems to have ridden roughshod over regulations in the hurry to build new AI data centers in Finland. However, the report suggests that it isn’t just environmentalists who have a beef with Google. Some families say wolves and other wild animals have entered the countryside villages. This would be a serious human safety concern.

“We had wolves coming into the village because all the wild animals moved out of the forest,” explained Anni Ranta, 53, whose family has lived in Otanmäki for generations. Ranta also fears the U.S.-owned data centers will make Finland an even juicier target for Russian aggression.

Finland is far from the first place targeted by AI data center developers to complain about environmental effects. This latest report also cites locals grumbling about construction/factory noise, streams turning brown, and rare birds being pushed out, among other unwanted side effects. We’ve seen similar concerns voiced across the U.S. and in other countries.

Google’s flouting of environmental policy has been unprecedented and clear, a representative of the Finnish Association for Nature Conservation (SLL) indicated to The Guardian. The Finnish Licensing and Supervision Authority (LVV) seems to agree, as it issued an order for Google to stop development near Muhos and Otanmäki. Nevertheless, a Google spokesperson was upbeat. “We will fully comply with LVV’s requests in this matter,” they said. “We look forward to maintaining an active, open dialogue with all authorities to strengthen our partnership with Finland and local organizations.” Google must be hoping the touted business benefits of the data centers will sway the government and/or regulators.

We last reported on these in-development data centers in Finland earlier this month, after the Finnish government (LVV) gave notice to Google to stop its construction activities in Muhos and Kajaani (near Otanmäki).

On this day in 2001, Apple CEO Steve Jobs unveiled the original iPod. He began the presentation by teasing “a breakthrough digital device that’s not a Mac.” Ultimately, the iPod that was revealed that day became a pocket-sized digital music player phenomenon, and would be many gadget fans' favorite always-carry device before smartphones matured enough to make such things redundant. The key claim iPod launched with was that it could put your entire music library in your pocket for $399.

In October 2001, when the iPod launched, Steve Jobs had successfully steered Apple beyond the beige into the Mac OS X era, with iBooks and iMacs well established. A revival of the Cupertino company's fortunes was well underway, with these iDevices doing well among consumers – at a difficult time, not long after the dot-com bubble burst.

Jobs predicted that the “Mac could become the center of your digital lifestyle.” It would do so as a hub uniting digital cameras, digital camcorders, DVD players – and the portable music player. Apple decided to go down this avenue as everyone loves music and it’s a huge target market, without a market leader, reasoned the CEO.

Small companies like Creative and Sonic Blue, and the much larger Sony, already had portable digital music players, acknowledged Jobs. However, “none of them has really found the recipe yet, for digital music,” he told the crowd at the iPod unveiling. At the time of Apple’s iPod announcement, there were four mainstream portable digital music solutions: CD players, MP3 CD players, flash players, and hard disk-based jukebox players. Jobs saw that last one as the optimal choice for market entry.

After this long preamble, the iPod was introduced as an MP3 music player, offering CD-quality sound, capable of playing all the popular formats of the time (MP3, WAV, AIFF). Its 5GB capacity meant that it could hold 1,000 songs. Specifically, that meant 1,000 typical pop-song length MP3s at 160Kbps. “For most people it’s their entire music library,” enthused Jobs.

The Apple CEO then sketched out three big claims that would make the iPod a winner. First, its ultra-portable form factor made it desirable. This aspect of the design was aided by Toshiba’s newest ultra-compact and ultra-thin hard drive at 1.8 inches in diameter and 0.2 inches thick. The actual device would be the size of a deck of cards – 2.4 x 4 x 0.78 inches, with a weight of 6.5 ounces, “lighter than most cell phones.”

Despite its compact size, the iPod offered advanced features like 20-minute skip protection, fast file transfers, and charging via a single FireWire 400 cable. The iPod was heralded as the “first and only music player with FireWire.” The connection was fast enough to transfer an entire CD to the device “in five to 10 seconds,” according to Jobs’ presentation. The USB connections of the same era would keep you waiting around five minutes for the same single CD of MP3s data transfer to complete, it was claimed. To fill the iPod (1,000 songs), it was estimated it would only take 10 minutes vs. five hours for a USB digital music player.

A rechargeable lithium-polymer battery in the first iPod offered 10 hours of continuous music playback. An empty battery could be recharged to 80% in one hour, Apple said.

Secondly, the iPod benefitted from “an amazing Apple design,” with minimal and durable back and sides, noted Jobs. It was touted to offer “Apple’s legendary ease of use” in a portable music player. Its specially designed controls were single-hand friendly, with an accelerated scroll wheel using a piezoelectric clicker for tactile feedback. It enabled the fast location and playback of music by playlist, artist, or song. Quiz fans – the first song demoed on stage was 'Building a Mystery' by Sarah MacLachlan.

The third breakthrough for the iPod claimed by Jobs was its Auto-Sync with iTunes digital music libraries feature. iTunes 2 launched on the same day with several advanced features like MP3 CD burning and syncing support for the new music player. However, the Auto-Sync meant that the libraries you created, customized, and curated on your Mac would always be synced with your pocket player when you connected to transfer music or charge. The iPod also doubled as a FireWire HDD.

The first iPhone launched over five years later, in early 2007. Both devices would sit alongside each other in the Apple product stack for over a decade. However, the iPod’s star inevitably faded, with the final model being the 7th-generation iPod Touch in May 2019. The product line was left unmaintained but was only officially discontinued in May 2022.

Blu-ray is fading after 20 years in the market, with Sony Japan finally discontinuing its last dedicated Blu-ray player. According to FlatpanelsHD, Sony Japan has finally marked the UBP-X700/K player as “Discontinued,” meaning customers in the country can no longer purchase a new standalone Blu-ray player from the company. This should not come as a surprise, as Sony stopped shipping the players in February of this year, suggesting that Sony only relied on existing inventory for its remaining Blu-ray player sales.

The PlayStation 5 is the only remaining Sony device that can play Blu-ray discs, but the company has also announced that it will end physical game disc production by January 2028. Because of this, some people fear that the upcoming PlayStation 6 will no longer have an optical drive. This move follows in the footsteps of LG, which ended Blu-ray player production in 2024. Japanese electronics companies Elecom and Buffalo have also ended production of Blu-ray drives for PCs this year, although Buffalo Americas says that it is “fully committed to supplying optical drives to the American market, and our DVD/Blu-ray drive products are anticipated to remain available for the near future.”

This move away from physical media and focus on streaming is worrying, especially to those who are concerned that we are moving towards a world where people no longer own anything. Because of this, sales of “retro media,” including CDs and vinyl, went up by 59% and 18% respectively. While you could still purchase a Blu-ray player from other manufacturers like Panasonic, Sony is one of the bigger companies that built and supported the standard, and its move to end the production and sale of Blu-ray players will leave this niche market even smaller. The audiophile market is facing the same dilemma, as some Hi-Fi player manufacturers ended production of selected products because optical drive makers have stopped making the components needed to read SACDs.

Ultra HD Blu-ray is facing the same upheaval that VHS and DVDs went through with the introduction of new formats. But while the older technologies were replaced by higher-quality formats, Blu-ray is being replaced not by more pixels, but by the convenience of streaming. However, it also comes with the caveat that people would no longer own physical media; a trend that worries data archivists.

TSMC, the world's largest contract chip manufacturer, is seeking alternative machinery without rare-earth components to reduce its dependence on China-dominated materials and avoid disruptions to factory operations caused by the country's export controls. According to a Reuters report on October 22 citing people familiar with the matter, the Taiwanese company has begun approaching its equipment suppliers to explore whether they can provide factory automation equipment equipped with motors that don't require rare-earth materials.

TSMC supplies the extremely powerful Nvidia chips that are facilitating the AI boom. It also manufactures chips for Apple, Broadcom, Qualcomm, and numerous other tech and automotive companies. The company operates enormous, highly automated semiconductor fabrication plants. These factories use industrial machinery that relies on electric motors with powerful permanent magnets made from rare-earth elements such as neodymium, dysprosium, and terbium.

According to the U.S. Geological Survey (USGS), China accounts for roughly 70% of global rare earth mining. The International Energy Agency (IEA) says China also controls around 90% of refining and approximately 95% of the finished permanent magnets found in TSMC’s machines, giving Beijing a quasi-absolute monopoly over one of the most critical technology supply chains. In April 2025, the country placed seven medium and heavy rare earths and the magnets made from them under export controls, requiring exporters to seek case-by-case approval and export licenses, thereby choking shipments to manufacturers across the U.S., Europe, and Asia, including TSMC.

China expanded the controls on October 9, 2025, to include five additional rare earths, processing technology and equipment, and even products manufactured outside the country that contain more than 0.1% Chinese-origin rare earths, or are made using Chinese processing technology. It later suspended the October expansion for one year, until November 10, 2026. This has now been extended for an additional 90 days.

TSMC’s moves aim to reduce the risk that these controls will hinder production at its factories. Even with all other raw materials and machinery required to manufacture its chips, shortages of specialized motor components could complicate the maintenance, repair, or expansion of its production equipment. Several technological approaches allow motors to avoid rare earth magnets. Conventional induction motors, for example, generate torque using electromagnetic induction rather than permanent magnets. Switched-reluctance motors are another option, while some permanent-magnet motors can use ferrite magnets instead of rare-earth magnets. However, these technologies typically involve different trade-offs in efficiency, size, power density, precision, and control requirements.

Meanwhile, the rest of the world is undertaking the painstaking, drawn-out task of building a rare-earth supply chain outside China. Taiwan is reportedly pursuing domestic rare-earth production capabilities in collaboration with the United States, with both countries confirming their intent to reduce reliance on Chinese rare-earth and critical-mineral supply chains.

Taiwan's government is supporting the Industrial Technology Research Institute (ITRI) in establishing a pilot production line within three years that could supply about 50% of Taiwan's domestic rare-earth demand. Meanwhile, the U.S. is “working with Taiwan and other partners to develop alternative sources of supply, expand processing capacity and build more resilient supply chains,” according to the report. In February 2026, the Trump administration announced plans to create a $12 billion critical minerals stockpile to offset supply chain risks from China.

Most everyone in the tech space has been living in interesting times, but this week isn't done yet, and it's already quite special for systems administrators worldwide. They already had to deal with the critical PHP RCE vulnerability, but they also have to update untold amounts of machines and software, thanks to eight insecure intermediate CA certificates that can allow attackers to impersonate any domain and/or pretend their malware is from a legitimate source.

An attacker anywhere in the path of your network traffic, like, say, in a company or government network, public Wi-Fi, your ISP, or one of any number of government entities, can intercept your connection and pretend that they're any website, like apple.com, amazon.com, or bankofamerica.com. If not updated, your operating system and installed software would be none the wiser. Additionally, with the code signing certificates, an application you download can be shown as completely legitimate from, say, Microsoft or Google.

The way that software checks a certificate is legitimate is by verifying its "chain." For example, a website like tomshardware.com has a certificate, issued and signed by an intermediate Certification Authority (or CA, a certificate-selling company), and whose certificate, in turn, is trusted and signed by root certificates — whose exceedingly precious keys are kept offline in safes. In this week's event, Indian CA firm eMudhra issued six TLS and two code-signing intermediate certificates with easily guessable keys.

Your operating system and potentially tens of different applications on your machine all keep a list of which root certificates and CAs are trusted. Browsers also keep a list of those not to trust, called a revocation list. These lists are updated periodically as screw-ups happen, just like the one that took place this week. And now they all need to be updated again, by revoking the eight CAs, removing the one affected root certificate, or just hard-coding a rejection of all eight affected keys.

There's a silver lining: recent versions of Chrome, Brave, Firefox, and Safari all enforce a feature called Certificate Transparency that consults a public read-only log of certificate issuances, and blocks any that aren't mentioned. Google reports that no certificates signed by the six affected intermediate CAs were logged, so if regular web browsing is all you did, you should be safe — but still check all your software for updates nonetheless.

The problem here is the sheer amount of machines and software that needs updating: PCs, servers, operating systems, phones, tablets, routers, web browsers, web browsers wrapped in apps, virtual servers, virtual machines, Docker and other containers, programming languages like Python, Go, Node.js, Rust, standalone download utilities like curl, and probably tens or hundreds of other categories. Additionally, we discussed things that are expected and likely to get updates. Old routers, smart and IoT appliances, and old computers could eventually present a problem.

As for the technical details, eMudhra issued six intermediate CA certificates and two intermediate code-signing certs with rather unadvisable private keys. The private key is the secret half of the two parts of a certificate, and for some reason, those of the aforementioned certs were only 1 byte long — thus a number from 0 to 255, trivial to brute-force in a split second. How the heck that happened is still under investigation.

Moscow is apparently using commercial fishing trawlers to map undersea cables connecting the U.S. and Europe. According to The Telegraph, it has uncovered over 500 incidents where Russian fishing boats loitered above critical communications cables over a span of four and a half years. All the data came from Global Fishing Watch, which uses satellite data to track vessels over the open ocean. The satellite tracks of these suspicious Russian ships are then cross-referenced over nautical charts that marked the locations of undersea cables.

While the general location of most undersea cables is clearly marked on nautical charts so ships can avoid them, any operation involving undersea cables requires more accurate positioning data. The Telegraph gave three examples of Russian trawlers seemingly surveying critical infrastructure. One ship followed the track of nine separate cables under the Atlantic, including the Atlantic Crossing 1 telecommunications cable and the decommissioned TAT-14 undersea cable owned by Sprint Corporation. Another ship loitered near the trans-Atlantic Hibernia Express cable and then followed the route of the Shefa-2 cable that connects Shetland to the Faroe Islands, while a third partially followed the tracks of the Amitie and Grace Hopper undersea cables, which connects the UK, France, and Spain to the U.S.

These ships have slowed multiple times, sometimes traveling at speeds between two and three knots. A senior naval officer said that the ships suspected of conducting these intelligence activities are large enough to house an uncrewed surface vehicle (USV) or remotely operated vehicle (ROV), and that they could deploy tethered units at very low speeds. “If you put a sonar-equipped ROV 100 meters above a feature of interest, you can map it with millimeter precision,” one naval officer said. A NATO representative also told the publication, “We’re absolutely aware of what Russian civilian fishing ships are up to and where ... Some of these ships have never seen a single fish on their deck.”

Undersea cables are the backbone of the modern interconnected world, and they carry economic activity worth billions, if not trillions of dollars. It’s estimated that around $87 billion worth of economic activity passes through this undersea infrastructure, and that they would be one of the first targets if ever a conflict between NATO and Russia erupts. These trawlers are part of Russia’s “gray-zone” warfare, where the actions of these commercial vessels cannot easily be traced back to the Kremlin.

There have already been multiple incidents of undersea cables getting disrupted across the world. Two undersea cables, one connecting Finland to Germany, and the other to Estonia, were disrupted one after the other in a span of a couple of months in 2024. Taiwan has also detained a Chinese freighter after it was suspected of severing an undersea cable connecting its Penghu Islands to the nation’s main islands, while another Chinese ship has been suspected of cutting an undersea cable connecting it to the U.S. More recently, NATO uncovered a Russian plot to sabotage the two largest undersea cables on Earth, while a cable connecting Australia and New Zealand to Indonesia and Singapore have also been damaged. Iran even threatened to sever undersea cables going through the Strait of Hormuz.

Even though undersea cables are considered critical infrastructure, the distances that they cover make it quite difficult, if not impossible, for nations to conduct constant patrols to protect them. The majority of their length also often sits in international waters, meaning vessels should be able to freely pass over and map undersea cables. While navies shadow these questionable vessels, their suspicious operation is generally not grounds for boarding, detention, or seizure.

OpenAI announced in a press release on Oct. 21 that researchers working under its “Patch the Planet” cybersecurity initiative used its AI model GPT-5.6 Sol to uncover critical security vulnerabilities in PHP (an open-source programming language that runs an estimated 70% of the internet). After the discovery, the company says it coordinated a coalition of researchers, PHP maintainers, security providers, and government cybersecurity agencies to fix the issues over two months. The company attributed the lengthy timeline to the task's complexity, which required major changes to the PHP engine.

The PHP Foundation has released patched versions PHP 8.2.35, 8.3.36, 8.4.27, and 8.5.12 to fix the vulnerabilities. PHP 8.1 and older versions will not receive official fixes. The coordinated release was originally scheduled for Oct. 22, but was brought forward to Oct. 21 following premature disclosures on Oct. 20, which led to public outcry.

OpenAI says every site using PHP can now safely upgrade its version to stay secure. The company highlights the incident — which it considers one of the biggest cybersecurity discoveries of 2026, given PHP’s ubiquity — as an example of the Defender's Window concept in action. OpenAI introduced the concept after its models hacked Hugging Face; it describes a critical, narrow timeframe during which organizations can use advanced frontier AI models to find, patch, and strengthen their cybersecurity defenses before those same AI-driven attack capabilities become widely available to malicious threat actors.

Operating as part of the Patch the Planet program, researchers from Calif, a California-based cybersecurity firm and the program's founding partner, began auditing PHP with OpenAI tools on July 30, 2026, at the PHP Foundation's request. The Foundation had known about the underlying flaw since 2024 but couldn't identify an exploitation method or a fix. Calif researcher Ryan A. first identified the exploitable vulnerability on Aug. 11, followed by OpenAI researchers that same day. Subsequently, another cybersecurity firm, Wiz, independently identified an additional vulnerability.

The two vulnerabilities, tracked as CVE-2026-109618 and CVE-2026-109619, involve flaws in how PHP manages computer memory and handles errors. The first occurs when PHP runs out of allocated memory while processing a request, potentially leaving behind incomplete internal data that the engine may later mistakenly treat as valid. The second involves a related problem in which a fatal error leaves PHP pointing to memory that has already been freed, allowing subsequent requests to access or corrupt it. Under certain conditions, attackers can exploit these weaknesses remotely — without needing login credentials — to execute malicious code on affected servers, potentially taking control of websites.

Using AI, the researchers also demonstrated that the flaws could expose sensitive information stored in server memory, including passwords, authentication tokens, and API keys. Most concerning, the vulnerabilities exist in PHP's core engine, meaning websites can be affected even if their application code has no security flaws. Both vulnerabilities received a critical severity score of 9.2 out of 10.

OpenAI says human developers and security teams examined the underlying code but couldn't fully resolve the flaws, allowing them to persist for two years after discovery. Meanwhile, AI-wielding researchers demonstrated practical ways to exploit it within two weeks. Conversely, one could argue that the flaws would never have been exploitable without AI, making them effectively harmless.

However, OpenAI is challenging this view with its Defender's Window concept. The company argues that powerful, revolutionary technology will inevitably be used by bad actors. It therefore maintains that it is a collective responsibility to act proactively to address any potential exploitable flaws. Anthropic's Claude Mythos recently identified 'thousands of zero-day vulnerabilities' in 'every major operating system and every major web browser.’

Google’s tensor processing units (TPUs) — its custom AI chips — now serve AI in production at Meta (creator of the widely used PyTorch AI framework) through a new PyTorch-native version of the SGLang inference engine, announced on Google’s developer blog. A matching version of vLLM, another widely used inference engine, was also announced as open source.

Previously, both engines used Google’s own JAX AI framework to run on TPUs. The new versions allow the hardware to be a native option for PyTorch, mostly without rewriting.

“SGLang-TorchTPU is running in production at Meta today” after “a close three-way collaboration between Meta, Google, and the SGLang team,” said Meta’s PyTorch-TPU lead Zongwei Zhou. That team works at RadixArk, a startup founded by SGLang’s creators, which built the engine with Google and Meta. The engine, he said, gives Meta “a remarkably mature, PyTorch-native serving stack.”

Jana van Greunen, director of PyTorch engineering at Meta, added in Google’s TorchTPU post that Meta “cannot be constrained by a single hardware architecture” and that TorchTPU “enables us to deploy our PyTorch workloads to Google TPUs.” Google said it co-developed TorchTPU with Meta’s PyTorch team.

This is not Meta’s first attempt at using TPUs. Google calls PyTorch/XLA, the older bridge from PyTorch to TPUs, a "predecessor" to the new TorchTPU, and SemiAnalysis reported last year that Meta research teams used TPUs through this bridge from 2020 to 2023 before Meta canceled those contracts. Google’s PyTorch-on-TPU push was aimed mainly at Meta, which did not want to migrate to JAX.

Although it is unclear where Meta’s production TPUs sit, talks last year reportedly included Google Cloud TPU rentals in 2026 with purchases in 2027, and Meta later reportedly signed a multi-year rental agreement.

The announcement came one day after Amazon said it was open-sourcing TorchNeuron, its own native PyTorch backend for its Trainium2 and Trainium3 chips, at the same PyTorch Conference. Both moves appear to be aimed at loosening Nvidia’s software lock-in.

On performance, a tuned vLLM setup serving Moonshot AI’s open-weight, 2.8-trillion-parameter Kimi K3 model reportedly tops 700 tokens per second for a single user on the seventh-generation Ironwood TPU. That speed figure comes from Inferact, the startup founded by vLLM’s creators that helped Google build vLLM-TorchTPU. Its test ran 16 Ironwood chips with speculative decoding, a method of speeding up large language model inference with a small draft model, against vLLM’s published setup on 16 Nvidia GB200s, which reached 452 tokens per second. Its kernels for Kimi K3 and Qwen 3.8 27B deliver roughly 1.4 to 2 times the GB200 baseline’s throughput at small batch sizes without speculative decoding.

SGLang-TorchTPU will keep adding models, starting with DeepSeek V4.1 Flash and Reflection Beam. It's also working toward Day 0 TPU support for new open models, meaning support on release day, by the first quarter of 2027, according to RadixArk. Google, Meta and RadixArk plan to share more at SGLang Summit 2026, Nov. 12–13 in San Francisco.

A computer, found hidden behind a fridge in a Japanese hotel room, was responsible for sending around 480,000 phishing emails, Hackread reports. The computer was placed there by a Chinese couple who lived locally, and the pair have now been arrested for both the hidden phishing PC and a separate slew of purchases that were made using stolen credit card details. Ye Jiafeng, 39, and his wife, Shen Qi, 39, now face multiple charges including fraudulent obstruction of business and trespass as well as computer fraud.

The scam began on 15 May, 2025, when Ye connected the computer to hotel Wi-Fi and then hid it behind a fridge. There it stayed, sending around 480,000 emails. When it was discovered, the machine had around 2.4 million email addresses, presumably just waiting to be phished. The emails claimed to be from credit card companies and sent unsuspecting recipients to fake websites in an attempt to capture their credit card information. It's believed that the computer was administered remotely, allowing it to remain in situ without any in-person maintenance required.

Ultimately, the computer would likely have remained in place if the recipient of one of the phishing emails hadn't reached out to the hotel. The hotel's IP address was found in metadata linked to the messages, and staff discovered the computer before reporting it to the police. Police are still investigating whether the two scammers were the brains of the operation or simply working for someone else.

While phishing was the couple's chosen scam this time around, it isn't their first foray into using technology to part victims from their cash. Ye and Shen are claimed to have used credit card details stored on a smartphone to buy items in Kobe. But that wasn't the only phone that they had — media reports claim that police found 74 of them at their home, and the couple had spent 10 million yen (around $63,000) of other people's money using stolen card details since 2024. It's unknown whether the phones found by police account for all of those that were used, or if there were others that have yet to be identified.

Phishing has long been a lucrative option for criminals around the globe, but it has also spawned a spin-off industry recently. A China-based phishing-as-a-service operation was dismantled by the FBI earlier this year. The service made it easier for criminals to create scam sites using AI and is believed to have been responsible for almost $2 billion in losses and 3.87 million stolen credit cards.

Intel on Wednesday announced that it had started installing production tools at its Fab 62 in Arizona, and the fab is on track for mass production by late 2028. The company also reported progress on its Ohio One manufacturing complex and confirmed that it has integrated High-NA EUV lithography into high-volume manufacturing (HVM). Meanwhile, Intel continues to invest in its leading-edge process technology R&D, advanced packaging, and photomask production capabilities in the U.S., which puts the company ahead of the competition.

"There is an important difference between technologies made in America and technologies that are truly American-made," wrote Naga Chandrasekaran, executive vice president and chief technology and operations officer, Intel Foundry. "The difference lies in the research and development process that transforms ideas into manufacturable technologies."

Fab 62 to begin operations in 2028

Intel completed construction of its Fab 62 shell some time ago; now it is time for the company to begin installing production tools, which is what it is doing now. Intel expects the new facility to become operational before the end of 2028, although it has not disclosed when it intends to begin high-volume production or which nodes it will produce.

Intel has not officially assigned Intel 14A to Fab 62, although the facility has been discussed as a potential manufacturing site for the node. Meanwhile, there is an important distinction between Intel's original plans for Fab 62 and its current commitments. When Intel announced Fab 52 and Fab 62 in Arizona, it said the two facilities would manufacture its most advanced process technologies, including Intel 20A (2nm-class) and 18A (1.8nm-class). Intel subsequently designated Fab 52 for Intel 18A manufacturing. Yet, Fab 62 remains unassigned.

Given Intel's manufacturing process roadmap disclosed earlier this year, Fab 62 will be able to make chips on 18A, 18A-P, and 14A nodes. Intel currently manufactures Intel 18A chips at the neighboring Fab 52, whereas Fab 62 could potentially accommodate either additional Intel 18A capacity or the company’s next-generation Intel 14A process technology. However, Intel has yet to confirm its plans for the facility.

The facility will become Intel's Fab 62, the sixth high-volume semiconductor manufacturing fab at its Ocotillo campus, which has served as one of the company’s primary production sites for decades.

Intel has also not revealed Fab 62's planned wafer processing capacity, or the amount of capital expenditures required to complete the facility. Nevertheless, the installation of production equipment indicates that the project has progressed beyond construction of the building itself. The new fab will expand Intel’s manufacturing footprint in Arizona, where the company already produces leading-edge processors and operates advanced packaging, assembly, and testing facilities.

Intel confirms High-NA EUV insertion into mass production

Another noteworthy disclosure concerns Intel's deployment of High-NA EUV lithography. Intel was the first chipmaker to install a commercial High-NA EUV scanner at its D1X fab near Hillsboro, Oregon, and subsequently integrated the tool into high-volume manufacturing to produce its Panther Lake and other chips.

The Gordon Moore Park at Ronler Acres in Oregon is where Intel develops its most advanced logic process technologies and initiates volume production of the latest processors. Usage of High-NA EUV to produce select tiles for Panther Lake CPUs at D1X is a natural fit for the development fab, as it helps Intel to better understand how these lithography systems work. Meanwhile, it seems (though Intel has not explicitly confirmed) that, for now, High-NA EUV scanners are used to produce the same circuits as Low-NA EUV machines, so the use of next-generation lithography systems does not necessarily provide significant advantages over existing litho scanners.

Ohio One reaches another construction milestone

Intel has also made progress with its 'Ohio One' manufacturing complex in New Albany, Ohio. Construction workers recently installed the first major structural truss at the site's first semiconductor fab, which represents a milestone in the development of the company's manufacturing operations in the state. The project is clearly an important part of Intel's long-term strategy to expand domestic production of chips using leading-edge process technologies for years to come.

Unfortunately, Intel did not provide an updated construction schedule for Ohio One or disclose when the first fab is expected to begin operations. Intel also did not reveal which process technology the facility will eventually use or the estimated manufacturing capacity it will provide. Therefore, while construction continues, which is good news, the announcement does not clarify exactly when Ohio One will contribute to Intel's output.

Other investments across the U.S.

In addition to leading-edge manufacturing capacity in the U.S., Intel continues to invest in advanced packaging technologies. Oregon serves as the company's main research and development center for advanced logic and 3D interconnect technologies, while Arizona hosts advanced packaging R&D, assembly, and testing operations. Intel uses its site in New Mexico to produce its complex multi-chiplet processors relying on advanced packaging methods — starting with modest EMIB and all the way to next-generation EMIB-T and Foveros technologies.

Also, Intel recently broke ground on a new photomask manufacturing facility in Santa Clara, California. The new facility will be an important part of Intel's supply chain in the U.S., as the company will be the only chipmaker in America to develop and produce chips using leading-edge nodes completely in the U.S. (both Samsung and TSMC currently lack leading-edge process R&D in America; they do not run their own photomask shops in the U.S., and Samsung has abandoned its plan to build an advanced packaging facility in Texas). Of course, Intel critically depends on tools from Europe and materials from Japan, but even as its major rivals invest hundreds of billions in their U.S. production capacity, Intel remains ahead.

To highlight its American supply chain, Intel disclosed that its U.S. operations involve more than 4,500 suppliers across 49 states, as its annual domestic expenditures exceed $29 billion. The company stressed that domestic semiconductor leadership requires not only wafer fabrication capacity but also research, process development, advanced packaging, and a comprehensive supply chain.

Elon Musk says ASML and other critical semiconductor equipment suppliers have expedited shipments for the ambitious Terafab semiconductor manufacturing facility, supporting his aggressive timeline for the joint megaproject between Tesla and SpaceX. He made the announcement during a live broadcast of SpaceX's Q3 2026 earnings call on X yesterday, October 20. Musk also said the company is targeting volume chip production at the project's primary facility in Gibbons Creek, Texas, by the third or fourth quarter of 2028, emphasizing that the timeline is highly aspirational and carries "big error bars."

"These are very aggressive dates, and we're very appreciative of ASML and all of our other suppliers in accelerating equipment deliveries in order to achieve this outcome," said Musk, CEO of both companies behind Terafab. SpaceX plans to manufacture its first experimental test chip at its research fabrication facility in Austin, Texas, around Q2 2027.

Aside from Terafab, Musk also outlined near-term ambitions for his AI companies during the broadcast, saying he expects them to have 2.5 to 3 gigawatts (GW) of computing capacity active by the end of 2026. He also stated he is highly confident about bringing an incremental 10 GW online in 2027, saying he "would be surprised" if they do not hit a total of 13 GW of AI compute by the end of 2027.

Meanwhile, the next-largest project by planned capacity, the now-restructured 10 GW Project Stargate, was expected to reach full capacity in the 2030s. Meta's 5 GW Hyperion also has a similar timeline. To achieve the 13 GW milestone, Musk revealed that SpaceX and xAI are aggressively procuring upcoming next-generation NVIDIA architecture. He announced plans to get a massive 440,000-GPU Nvidia "Vera Rubin" training cluster online by the end of Q2 2027.

Shifting back to Terafab, Musk also boasted that once fully scaled and operational, the sheer commercial value of Terafab's advanced chip output will place it within the top 20 global country GDPs from a single physical building. For context, the 20th country on the list is Switzerland, with a GDP of around $1.14 trillion. The facility is a planned multi-billion-dollar, vertically integrated chip factory in Texas, with an ambitious goal of manufacturing 1 terawatt (TW) of AI computing power per year. Terafab will fabricate logic, memory, and advanced packaging internal architectures in one place. The project's massive planned output has mandated various collaborations across the semiconductor supply chain. Terafab will utilize traditional extreme ultraviolet (EUV) lithography systems utilizing Intel's 14A manufacturing node technology.

Intel officially joined the Terafab venture in April 2026 to help co-develop and refactor silicon fabrication technology, providing access to its state-of-the-art 14A manufacturing process. ASML, the world's sole commercial supplier of the $400 million EUV lithography systems required for high-volume, leading-edge chip production, is perhaps the project's most critical partner. Terafab's first EUV tool was reportedly slated for delivery in mid-2027, according to unofficial sources. It remains unclear if this is the expedited timeline Musk referred to in the broadcast.

Other major suppliers include American equipment manufacturers Applied Materials, Lam Research, and KLA Corporation, as well as Japan's Tokyo Electron. These companies will provide front-end processing, metrology, and process control machinery. Tesla and SpaceX are also in talks with about 10 South Korean equipment makers, including Hanmi Semiconductor, known for advanced packaging equipment, and HPSP, a high-pressure hydrogen annealing equipment manufacturer, as well as firms providing specialized piping components and process valves.

Musk's announcement says that several of these suppliers are expediting supply to Terafab, even as capacity remains tight across the semiconductor supply chain. The facility, which aims to mitigate future shortfalls of cutting-edge AI processors from external foundries like TSMC, will supply the chips for Musk's xAI clusters, self-driving vehicles, Optimus humanoid robots, and SpaceX's orbital AI data center, as well as the broader AI industry.

However, many experts have questioned the validity of Musk's 1 TW Terafab ambition. For context, research firm Gartner estimates the total global AI-optimized data center capacity is roughly 44 GW as of late 2026, 4.4% of Terafab's planned output. Although global AI computing capacity has grown by roughly 3.3x to 3.4x per year since 2022, according to EpochAI.

President Donald Trump has issued a call for data center builders and AI hyperscalers to pay communities a dividend in order to overcome energy-related price hikes, or even to reduce prices, in order to make data centers more palatable to local communities. In a post on Truth Social post, he warned that data centers would be mandated not only to make their own electricity, as he has previously stated, but also to give or sell excess energy back to the local community.

Local communities all over the United States have protested against AI data centers, often owing to their outsized impact on local communities. Rising electricity costs, noise pollution, water shortages, air pollution, deforestation, and more are the things that small communities have to deal with because of these companies.

To ease the pain of local communities, Trump is demanding that these large companies provide the communities they affect with something that will hopefully make their presence more acceptable. The Trump Administration is keen to stay ahead of China in the AI arms race and is looking for solutions to win back some favor by making the data centers pay something back to the affected communities; how that will practically transpire is yet to be seen. Considering the electricity-focused post, the dividend might be in the form of partially paying rising electricity costs to alleviate a pain point that directly affects many homes within the vicinity of data centers. The Trump Administration had ordered these companies to make their own electricity, as we’ve covered previously. Excess electricity from their generation could be used to further lower the electricity cost for the locals, including a recently passed act that orders data centers to shoulder the cost for grid upgrades to avoid taking it from the taxes of local communities.

“Data Centers that were highly unpopular in a certain area have become extremely popular when a financial deal is struck.” Trump further says in his post, aiming to have more data centers built, “MAKE DEALS, and get them built!!!," the President said in closing.

Whether the dividends will be enough to win over communities is yet to be known, as the economic aspect is only one facet of the problem. There are still plenty of problems to face; enormous water consumption, ecological damage, and pollution are still major challenges that make AI data centers a hard sell anywhere they’re located.

Qualcomm has told employees it will cut jobs worldwide after its annual performance reviews, and it expects to begin informing affected employees in November. According to a Reuters report, CEO Cristiano Amon framed the move as a “performance-focused reduction” intended to build a “stronger performance culture,” stating that the company will consider organizational needs, business priorities, and employees’ skills and expertise when determining affected positions.

“The reduction will address employees who have consistently fallen short of expectations,” said Amon. The memo did not specify which departments would be affected or how many jobs would be cut. The company had roughly 52,000 employees at the end of its 2025 fiscal year, typically in late September.

Qualcomm, one of the world’s largest smartphone chip manufacturers with customers including Apple, Xiaomi, and Samsung, has been hit by an industry-wide decline in consumer electronics sales as rising memory costs reduce demand. Its shares have risen just 1.5% in 2026, compared to gains of almost 60% for the Philadelphia Semiconductor Index.

As a result, Qualcomm is strategically pivoting to AI infrastructure and edge computing, following a broadening trend across the semiconductor industry. The company unveiled its AI200 and AI250 AI inference accelerators in October 2025. After its latest earnings report projected lower quarterly profits than initially estimated, the company said that growth in its data center business would offset any lost revenue by fiscal year 2027.

Qualcomm’s layoff decision comes as it tries to position itself for future growth. While mobile chips remain its core business, its best future opportunities might lie in data center silicon, automotive platforms, personal computing, and edge computing, and all those fields require specialized engineering talent. A performance-based reduction could let the company retain its most valuable employees in those fields while cutting duplicated roles, lower-performing teams, and positions that no longer fit its strategy.

"As we have expanded the business, so have the competencies ⁠and skills ​we require," Amon wrote. "We must have the right ​people in the right roles so they can perform at their best and the company can continue ​to succeed." The move comes on the back of Q3 2026 financial results that included a 4% year-on-year decline in revenue and 25% lower net income. The core handset business recorded a 20% decline in revenue on the year, too. The company reportedly recorded $97 million in severance-related restructuring charges in the first nine months of its fiscal year.

Qualcomm’s move is part of a larger industry transition, largely driven by AI. After several years of aggressive hiring, many tech companies are now trimming their workforce to fund AI ambitions. Meta announced plans to cut 8,000 jobs to fund AI infrastructure. Even as companies are cutting jobs that don’t fit their AI roadmap, they're fiercely competing for process engineers, chip designers, and AI software developers. McKinsey and the SEMI Foundation estimate that the United States could face a shortage of up to 157,000 semiconductor workers by 2030, while 73% of chip companies report difficulty filling engineering roles.

The Islamic Republic of Iran has described several undersea fiber cables that run through the Strait of Hormuz as "legitimate new targets." The area has become a hotbed of conflict since the start of military action by the United States, Israel and others in February. General Hossein Mohebbi, a spokesman for Iran's Islamic Revolutionary Guard Corps, referred to the cables in comments at an event this week, The Telegraph reports.

"Seven large cables pass beneath the Strait of Hormuz, and all internet and commercial data pass through them," according to Mohebbi. He went on to describe the cables as "legitimate new targets" for future assaults. Meanwhile, another IRGC General, Yadollah Javani, who serves as the IRGC's political deputy, has since suggested that if Iranian infrastructure is attacked, Iran will respond by attacking the cables, according to comments reported by Iran International.

The conflict between the United States and Iran has cooled, but it isn't quite over yet. Mohebbi suggested that a "third phase" in the war would reveal a "new geography" that would include "new targets and new weapons." One particular cable referenced by the IRGC appears to match the Asia-Africa-Europe-1 cable, a cable stretching 15,050 miles from Asia to Europe, traveling through Oman, the UAE and Qatar. Both the UAE and Qatar share direct maritime borders with Iran.

Experts believe the threat is credible. Professor Alesso Patalano, a war and strategy professor at King's College London, told The Telegraph that "not only is it possible, but it's been a rather clear possibility for a long time," adding that divers could "go down to the seabed... and cut them or destroy them." Such attacks aren't "unheard of," going on to reference the Nord Stream 2 gas pipeline attack near Denmark in 2022, supposedly by Ukrainian agents, as an example.

Meanwhile, further comments, made to The Telegraph, suggest that Iran had considered cutting the cables several times since the war began. However, up until now, Tehran had avoided doing so, as it would mean "declaring war on the whole world and building new enemies," according to another unnamed Iranian spokesman, although stating that it would continue to be "an option if Trump escalates."

Iran isn't the only country that has taken a special interest in undersea infrastructure. A Russian ship, for instance, was chased from British waters by a Royal Navy attack helicopter earlier this year as it loitered for 14 hours less than a mile from five undersea cables that connect the UK with Spain, Portugal, Canada, and the U.S.

A fanless compute module will head to the Moon for supercomputer benchmarking, AI testing, and additional experimentation by NASA and Deloitte, as detailed in a Hewlett Packard Enterprise (HPE) release. The company’s Spaceborne Computer-4 will run what HPE calls “the first high performance computing (HPC) benchmarks on the surface of the Moon,” riding on an Astrolab FLIP rover to the lunar south pole. The launch is slated for late 2026 under NASA’s Commercial Lunar Payload Services initiative for a Moon Base program mission, in conjunction with a Voyager Technologies Griffin-1 lander.

Named benchmarks include High Performance Conjugate Gradients (HPCG), a supercomputing metric that complements High Performance LINPACK (HPL). AI testing will involve large language model (LLM) retrieval-augmented generation, vision inference, LiDAR processing, and an AI services container for AI app deployment. The LLMs “will deliver insights, analyze visual data, and support autonomous monitoring,” while the LiDAR processing involves mapping, navigation, terrain analysis, and infrastructure development.

Scott Wiest, HPE’s vice president, fellow and global sales chief technologist, said, “The next chapter of lunar exploration will depend on the ability to turn data into insights at the point of discovery.” HPE says “no other company or organization” has conducted such benchmarks on the Moon, with a presumption that no other parties’ planned HPC missions will get there first. HPE also assembled a software suite that handles the link to Earth-side systems and the secure delivery of results.

HPE’s “lightweight, self-contained, and fanless” advanced compute module is “built for space environments,” bringing AI and HPC “to the lunar edge,” as Wiest put it, or the “far-edge,” as HPE calls it. The computer blade is about 10 inches long and about 4 inches wide, about the size of a typical graphics card, and is made from two component cards, as HPE senior director Norm Follett explained on the company’s Technology Now podcast last year.

It includes 1TB of Kioxia storage, a low-power CPU, and a small GPU. The chassis is designed to shed heat into a rover plate, and the overall design is made for the Moon rather than being an off-the-shelf solution like International Space Station (ISS) units. Currently, HPE’s description of the unit lacks hard details on performance targets, power budgets, or chip vendors, but the eventual benchmark results will yield some information.

The original Spaceborne Computer launched in August 2017 on SpaceX’s CRS-12 to the ISS. It also ran HPL with HPCG, surpassing one teraflop, according to HPE. It was rack-installed and water-cooled; the fanless Computer-4 is quite miniaturized in comparison to the 124-pound original Spaceborne Computer. Spaceborne Computer-2, also on the ISS, is made of HPE Edgeline and ProLiant systems with Kioxia SSDs.

NASA will use the opportunity to run its Federated Learning for Unbounded and Intelligent Decentralization (FLUID) AI neural network, trained on data from the Earth, the Moon, and the ISS, coordinating three distinct models. The agency will also test a tuned LLM for semantic processing for handling onboard documents. Deloitte, for its part, will test its Silent Shield, “a new system to detect cyber intrusions” with near-real-time predictive analytics. Testing ideally occurs onboard, as the link to Earth is slow with low levels of bandwidth, Follett explained.

The launch is planned for the fourth quarter of 2026, using a SpaceX Falcon Heavy launch vehicle, SpacePolicyOnline reports. It will take several days for the lander to reach the Moon and its surface. The FLIP rover is designed to operate for one lunar day, or about 14 Earth days, before attempting to survive the night via hibernation, Follett said, making for a relatively tight benchmarking window.

Three contract workers died, and one more is in intensive care after an accident at the construction site for TSMC’s 2nm Fab 22 in Kaohsiung, Taiwan, reports Taiwan’s Central News Agency. The Southern Taiwan Science Park (STSP) Bureau suspects the cause was a broken bolt on a steel connection plate that sent the workers and steel components tumbling. Steel hoisting on the project has been halted, and the STSP has sent labor inspectors to investigate. The accident is not the first of its kind but rather follows at least a half dozen others since September 2024 at TSMC construction sites in Taiwan, according to local reports. Each of those earlier accidents killed one worker.

Local police said that the men were welding a steel beam on an upper floor and fell with it, while the STSP said the workers were performing steel-structure hoisting and installation when a third-floor connection-plate bolt apparently broke.

10 emergency vehicles were sent to the site with 22 personnel, and all four men involved in the incident were taken to the hospital. On-site medical staff gave first aid, but three of the workers died despite resuscitation efforts, and the fourth was moved to intensive care. The workers were employed by a contractor working on the fourth phase of the complex’s construction, and TSMC said the area was within that contractor’s scope of work.

According to the report, TSMC stopped the related construction work on its own and is working with authorities and supervising the subcontractor’s follow-up.

The Fab 22 complex, located in Kaohsiung’s Nanzi District, is part of the science park and is being built to accelerate 2nm manufacturing. TSMC entered high-volume 2nm, or N2, manufacturing in the fourth quarter of 2025, including at Fab 22. The complex has five phases planned, with a total investment TSMC puts at more than NT$1.5 trillion, or about $46 billion. The accident happened on the construction side, not in a producing fab. Kaohsiung’s deputy mayor said construction on Phase 4 began in August 2025, with operations expected in the fourth quarter of 2027.

Business Today reports that TSMC fab construction has increased, from four fabs a year in 2017–2024 to nine a year in 2025 and 2026. The deaths come as the company plans for 13 new leading-edge and advanced-packaging fabs in Taiwan, with company-wide capex of $60 billion to $64 billion this year.

The faster pace has encouraged the addition of a contractor’s safety record as a weight in selection and a plan for a Taiwan Semiconductor Construction Engineering Safety Association with construction partners. TSMC Chairman C.C. Wei said at last year’s shareholders meeting that safety cannot be compromised, and the company’s 2025 annual report said it will keep investing in fabs across several locations in Taiwan “with a strong emphasis on construction site safety.”

Cliff Hou, TSMC senior vice president and deputy co-COO, told Business Today that increased worker count and overlapping construction have raised the risk, requiring careful supervision, and that TSMC’s contractor lost-time injury rate had fallen about 42% from last year’s high as of late May, when the publication toured the Nanzi site to see the overhaul. It’s too early to call the program a failure, with the cause as yet unknown.

The accident comes less than a week after an explosion killed two people at Intel’s chip plant in Israel, though that was an operating fab rather than a construction site. Both accidents come as chipmakers race to meet demand, with global chip sales passing $1 trillion through August, according to the Semiconductor Industry Association.

The STSP’s findings will determine what penalties, if any, will fall under Taiwan’s Occupational Safety and Health Act, while it assists the victims’ families. No date has been given at this time for that or for the Phase 4 steel hoisting restart. Nor has TSMC commented on any schedule impact.

Protesters disrupted Jensen Huang's Nvidia GTC keynote in Berlin on Wednesday morning. Just minutes into the CEO's customary address, the broadcast was interrupted by a heckler from the crowd demanding the CEO answer claims that the company's chips were found inside Russian drones, and charges of environmental destruction more broadly. While the incident wasn't fully picked up by Nvidia's livestream, footage filmed by attendees shows protestors standing on chairs with a large banner that reads "Nvidia, clean up your dirty business!"

Multiple videos and images shared to X show protestors brandishing a large banner with a Greenpeace logo. "Jensen, what about the technology that was found in Russian drones? Those were Nvidia chips. What are you doing about that?" one of the protestors called out across the auditorium. "What are you doing about that? What are you doing about the environmental destruction?" the heckler demanded.

Huang can be heard on the broadcast and in the video stating, "It's okay, go easy," as security rushed to escort the protesters out of the hall. "I'll talk to you later," Huang stated as the individuals were removed from the premises.

The complaint raised by the protester appears to relate to recent reports that Nvidia Jetson chips have been found in Russian cruise missiles used against Ukraine, as well as in AI drones. Reports indicate Russia's S-71 'Monochrome' weapon has autonomous targeting capability powered by Nvidia's consumer-grade Jetson Orin chip. "Our Jetson Orin modules are consumer-grade products sold to students, developers, and startups for a wide range of beneficial applications," Nvidia told Tom's Hardware at the time. The company reiterated that the chips are not available in Russia and are not designed for military purposes, but admitted it cannot track the resale of pre-owned Jetson chips after they are sold.

A report in August claimed that a Russian AI drone killed three civilians in Ukraine, the first documented case of civilian deaths caused by a Russian drone using fully autonomous targeting. Images of the recovered drones confirmed they contained Jetson Orin modules. "All important issues, but let me get through my keynote," Huang quipped before continuing his presentation.

Huang is in Berlin for Nvidia's three-day GTC event, taking place from October 20-22. The customary keynote, given this morning, also includes fireside chats and talks for developers and business leaders.

The U.S. Army is allegedly developing a program to deploy autonomous AI agents that can navigate digital and network infrastructures, identify targets, and “neutralize” them. According to an Oct. 19 report by The Intercept, citing leaked official documents from the Army Cyber Command (ARCYBER), the nascent program, codenamed “ECHOTRIBBLE,” is primarily focused on offensive missions with the goal of neutralizing targets by “disabling the threat completely or temporarily, deleting the threat from the target.”

The program purportedly seeks to develop modular, autonomous AI agents that would function like computer worms, stealthily navigating adversary networks, moving between devices, and self-replicating across different operating systems. The agents would operate lightly in contested cloud environments and air-gapped systems alike, keeping reports minimal and concealing any required transmissions. Once an agent identifies and neutralizes a target, it would self-destruct, ideally without alerting the target or leaving any traces of its existence.

Cybersecurity experts have raised safety concerns about the project. A recent string of events has seen AI agents escape their environments and autonomously execute unassigned tasks. The document says the army will implement “predefined triggers” to help agents precisely distinguish targets. However, experts worry that if agents designed for benign tasks could execute unplanned, sometimes dangerous, actions to achieve those tasks, what should be expected of self-replicating agents designed specifically for offensive missions? Anthropic’s experiments showed AI agents can set up inference on different hardware and replicate across a network.

The Army also wants the agents to be capable of “Self-Directed Behavior,” acting with as little human oversight as possible, as this is necessary to achieve the desired level of autonomy. Conversely, this level of autonomy has seen AI agents bypass guardrails to complete their tasks. Government intelligence agencies have also been known to penetrate target networks through civilian infrastructure. “The target may well be hit. But what path will it take?” asked Hyrum Anderson, co-founder and CEO of the AI security firm Abundant, as quoted by The Intercept. “Will it go through a U.S. water utility on its way, because that’s where the adversary has staged its command and control, and take that down too?”

Another angle of concern is how the agents would make judgment calls. While AI agents are excellent at overcoming obstacles in their path, there's no predicting with 100% accuracy what they would consider obstacles or what path they would take to overcome them. “One of the reasons that we trust 19-year-old Marine corporals in Ramadi with live rounds and a weapon is because they have both the compute substrate — the brain and the eyes — but the deep world knowledge to understand, ‘Yeah, I probably shouldn’t do that,’” said William Marcellino, a senior behavioral scientist at the RAND Corporation researching military AI applications. “That is not something that current public AI technology has.”

The ECHOTRIBBLE program is in the early prototyping and industry assessment phase and is led by ARCYBER, the Army Cyber Technology and Innovation Center (ArCTIC), and the Cyber Fusion Innovation Center (CFIC), with private vendors and academic researchers proposing and designing the software. ARCYBER and ArCTIC dictate the operational requirements and rules of engagement, while CFIC, a defense non-profit organization, acts as an intermediary between the Army and commercial tech companies. The Army recently held a virtual Assessment Event from October 6 to October 8, 2026, where it evaluated 65 technical whitepapers submitted by private vendors proposing solutions for the agentic framework and narrowed them to five. The final candidate(s) will be issued funding and formal prototype contracts.

The current and expected power draw of AI data centers has been the subject of strong discussion, but today, there appears to be some good news. Three U.S. nuclear reactors are being restarted after many years offline, expected to bring 2.25 GW total to the grid. According to a report, the Palisades Nuclear Plant in Michigan, the Crane Clean Energy Center on Three Mile Island (Pennsylvania), and the Duane Arnold Energy Center in Iowa are getting a fresh set of proverbial tires to get rolling again.

The Palisades plant was under active decommissioning, apparently marking this event as the first time a nuclear reactor is being put back to work after dismantling had begun. Despite what sounds like a tough starting point, this plant is the one that's furthest along, as it was loading fuel in August 2026 and is expected to generate 800 MW around March 2027.

Much like the other projects mentioned here, the Holtec Palisades firm running the reactor took a loan from the U.S. Department of Energy. Palisades borrowed from the DoE under the Biden administration's 2024 investment in nuclear energy, while the others were issued under the purview of the Trump administration's 2025 push.

The second reactor is the Crane Clean Energy Center on Pennsylvania's Three Mile Island, repurposing the facility's Unit 1. The project started in 2024, with Constellation Energy saying it would spend $1.6 billion total for the reactor. The firm then took a $1 billion loan from the Department of Energy in 2025. Microsoft is set to be the main customer, with a 20-year contract. This facility is projected to bring 835 MW online sometime in 2027.

This facility was famous in 1979 for the partial meltdown of its Unit 2, although mercifully the event had relatively tame environmental consequences. Constellation's ongoing project is restoring and refurbishing Unit 1, which didn't show significant issues until its 2019 decommissioning and has undergone a safety inspection from the Nuclear Regulatory Commission. This entity still needs to grant its blessing before the reactor starts loading fuel.

The Duane Arnold Energy Center in Iowa is the third unit being brought back to the grid, in a collaboration between NextEra Energy and Google. The predicted output is 615 MW, with the juice expected to start flowing in 2029. Next Energy CEO John Ketchum stated that the firm was happy with its cooperation contract with Google, as it meant that "none of the costs are being paid by residential customers in Iowa."

All three facilities had previously been shut down at various points as they were deemed too expensive to run, as labor and maintenance costs made nuclear-generated watts pricier than alternative means of production.

But the financial landscape for power has shifted, plus with such overwhelming demand and grid connections being a primary blocker for data center buildouts, their operators are likely willing to pay for premium power as long as they get it now instead of in a few decades.

Some quick browsing indicates that Michigan is going to need somewhere between 1.2 and 2.3 GW for data centers alone around 2030. Pennsylvania's forecast has a wide range of 0.8 to 3.4 GW, while Iowa's range is 1.55 to 2.8 GW. These figures are loose, as they include planned capacity for data center projects that may or may not end up being built, and whose timetables may slip due to economic conditions, hardware availability, and public resistance against expansive buildouts.

Modern computing systems and devices typically require pieces of crucial open-source software, typically maintained by unpaid or low-paid volunteers. Recently, a report collated a list of critical open-source projects, noting which of them are chronically underfunded and/or understaffed. The information is sobering: 11 out of 23 of the listed projects currently only have one or two maintainers, and funding is generally scarce and sometimes nonexistent.

The Data Drop's report on open-source maintainers identifies that having only one or two sets of eyes on critical pieces of software means that sooner or later something breaks — and it tends to do so in spectacular fashion, given its ubiquity. So, what happens when these projects, which are widely used across software, go underfunded or understaffed?

Lessons from the past

In 2014, the same year as Heartbleed, Shellshock made waves: a vulnerability in the Bash shell that's present in Linux, macOS, and Unix-type systems. The problem was traced back to a line of code present since 1989, displaying how even software that's thought to have all the bugs hammered out of it can still fall prey to attacks. A quick look at Bash's commit history shows it's almost exclusively maintained by Chet Ramey.

In 2016, ImageMagick was exploited, providing attackers with remote code execution. This library is used for image manipulation across a wide range of software, in webpages, mobile apps, and more. A 2020 analysis revealed that hundreds of security issues had been found over the years in its extensive codebase. Nearly every code contribution in the past couple of years has been done by an individual named Dirk Lemstra.

Contemporary examples include 2021's Log4Shell exploit, a critical vulnerability in the Apache Log4j logging library used in millions of servers. The libwebp image-handling library saw a zero-day vulnerability in 2023, and the xz compression library used across Linux and iOS was compromised in 2024.

Thankless work

Key points in the report mention specific repos, which might be under threat. Somewhat unsurprisingly, the low number of overall maintainers and the unpaid nature of the work make it unattractive to perform.

The main time-zone file used across Linux-based operating systems is maintained pro bono by Paul Eggert, a UCLA teacher. Core-js, a piece of support software that lets new JavaScript work in old browsers and is estimated by the author to be used in half of the busiest websites in the world, is handled by Denis Pushkarev. He once had to endure criticism from the community because he added a command-line message to the library's installer asking for help or a job.

The HarfBuzz library developed by Behdad Esfahbod is a cornerstone in ligatures and text shaping in Middle Eastern languages. In those writings, a letter changes shape and/or orientation according to what's around it. Esfahbod handles that vast majority of the project's code, and doesn't mention public or corporate funding.

Libjpeg-turbo handles JPEG images on Android, Firefox, LibreOffice, and every Chromium-based browser. It's kept ship-shape by D.R. Commander aka DRC, who occasionally is contracted for library-specific work, though at arguably modest values of "3-4 figures rather than 6 or 7." Libxml2, just like the label says, handles XML reading and writing across numerous applications and operating systems. Nick Wellnhofer, the original maintainer, stepped down in September 2025, after years of work with little income.

The SQLite project provides a one-file database that's exceedingly easy to integrate into most software, instead of complicated, resource-heavy databases, while still retaining standard SQL language and accouterments. Thanks to its ease of use, it found its way into an astonishing number of applications. Examples include WhatsApp, Dropbox, Android, Windows, and all major browsers.

The project estimates that there are about a trillion instances of SQLite in use. Just four people keep this critical project together, and they've gone to the length of removing their names and photos from the project's page, as "there are a lot of crazy people in the world who might misuse that information." The project is funded by the offering paid support via the founder's company, though there's no word on effective income, especially when it presumably needs to feed four mouths.

Funding is there, but scarce

It's not all doom and gloom, though. Take curl (the software), for example. Almost every developer has used it, thanks to its URL-fetching abilities and extensive scripting options. The project was started by Daniel Stenberg, who is now paid to work full-time on it. The team has now grown to 11 and has received funding through Open Collective and Germany's Sovereign Tech Agency, among others.

Likely due to the lessons learned from their past respective debacles, both the Log4j and OpenSSL projects are likewise funded by the Sovereign Tech Agency, as are the FFmpeg video file handler and the OpenSSH secure shell software that's ubiquitous in Linux servers.

All things considered, it's easy to argue that the critical importance of all the aforementioned software merits sizable investments from industry players, especially those whose commercial products benefit the most from the community's efforts. Plus, even if the firms feel no moral obligation to contribute, in the day and age where AI agents are finding exploitable vulnerabilities in hours, donations to these open-source projects could also be considered investments.

Intel has asked the Federal Communications Commission (FCC) to make an exception for the automated guided vehicles (AGVs) that the company uses to transport the sensitive silicon wafers and other materials, containers, and supplies needed to make the most advanced chips in its fabs. According to Bloomberg, the American chipmaker fears that its AGVs would fall under the blanket prohibition against foreign-made robots after they were added to the FCC’s Covered List.

AGVs are niche machines that chip fabs use to transport materials and supplies between tools and sections of the fab, helping ensure that the cleanroom conditions required for each manufacturing process are maintained. This makes them critical equipment for U.S. semiconductor manufacturing, but, apparently, there are no suitable domestic alternatives for these devices at the moment. Intel’s current fleet of AGVs requires continuous maintenance, including software and firmware updates, and if it’s forced to find a local supplier for its AGVs, the company says that it would take years, crippling the United States’ chipmaking efforts.

The biggest AGV suppliers are Japanese companies like Murata Machinery and Daifuku Co., while a few smaller businesses from South Korea and Taiwan are also trying to break into the market. It’s unclear why these companies haven’t applied for a conditional approval yet, which would allow them to continue selling these products in the U.S. while also promising Washington that they’ll set up a manufacturing base in the country. This is something that Netgear has already done, allowing it to continue selling its foreign-made routers despite being an American company.

Intel says that it remains unclear if updates to its AGVs would break compliance with the FCC, and that it’s asking the agency to issue a conditional waiver “out of an abundance of caution to ensure the ongoing operation of these specific AGV units.” While Team Blue is the only company concerned about AGVs at the moment, other chip manufacturers would likely need a similar waiver in the near future, especially as they build more fabs in the U.S. This includes TSMC, which is planning to build 12 more fabs and four packaging facilities in Arizona, Samsung, with its new Taylor, Texas, fab, and even Elon Musk’s Terafab, which has already begun construction.

Washington has been using the FCC to force companies to invest in American manufacturing. Its blanket bans on foreign-made products like drones and routers have affected some of the biggest brands in the U.S., including DJI and TP-Link. While these bans do not outright ban the sale of all products made by these companies, especially those that have already been approved before, they make it impossible for them to introduce new products and even release software updates to their existing product line. The FCC eventually walked back on the software update ban, but they still would not be able to register new products unless they build them in the U.S.

Millions of dollars worth of gift cards were stolen and then used to buy high-value tech gear at retailers like Target, Sephora, and Lowe's after a Chinese crime ring gained access to an internal Staples system and then used it to issue more than 15,000 cards to itself. While the hack hasn't been made public to date, a 404 Media report notes that it actually took place back in 2023.

The outlet reviewed court records that show that Homeland Security Investigations (HSI) was told about the hack in October of that year. The hack came to light after it led to a search warrant that was unsealed last week. And while the document doesn't specifically name Staples as the victim of the hack, 404 Media has confirmed that it was the company affected based on other information. Such information includes the fact that the victim company is headquartered in Massachusetts, that it uses Blackhawk Networks to fulfill gift card orders, and that the company uses an internal customer support system called Nexus. The report notes that a reference to Nexus can be found on a former Staples executive's LinkedIn profile.

While Staples was under the impression that Nexus could not be accessed from beyond its own network, hackers found that not to be the case. They then "performed multiple heap dump requests," a process that captures all of the data held in an app's memory. That information gave the hackers the tools they needed to hijack logged-in sessions of Staples customer service employees. From there, they were able to issue gift cards at will.

Once the gift cards had been issued, the crime group sent runners to physical retail stores to spend them before they were identified and canceled. Target found that 35 gift cards were used at a New Hampshire store, with high-value products from Apple, Bose, and Beats all bought. All 35 cards were used within a period of just a single hour, although not all of them were obtained via this hack.

In total, Target told investigators that 3,850 gift cards worth over $140,000 were used across 85 different stores in just two days. Sephora was hit for $58,600 worth of sales, while Lowe's saw $226,000 spent across online and in-store purchases. In total, 4,630 stolen gift cards were used with a value of around $425,000.

This hack was just the tip of the iceberg, too. Last April, a Department of Justice announcement confirmed that three Chinese nationals had been sentenced for their part in a gift card fraud conspiracy. Including the Staples hack, the crime group stole more than $100 million worth of gift cards.

Chinese organized crime groups often pop up on the radar of U.S.-based law enforcement. Earlier this year, a Chinese phishing service that was used to generate scam websites using AI was dismantled. The service was linked to almost $2 billion in losses.

The game of Minecraft is no match for a team of four Gemini 4 Argon agents, as they were able to defeat it in 27 hours with 2,783 self-written programs, a paper from Google DeepMind reveals. The agents did die 605 times, even on Easy, and about half the run was spent fighting the Ender Dragon, the final boss of the game. The run finally came to completion when the End was solved by stacking dirt to reach the boss’s crystals, at an estimated cost of around $520 in tokens.

The experimental run was designed from the get-go with one goal: kill the Ender Dragon. The playthrough involved a random, shared survival world on Easy with four agents — a fifth agent existed only for reporting — from Gemini 4 Argon, Google’s latest frontier model. Each agent had its own player through the open-source Mineflayer JavaScript library for Minecraft bots, but they were able to coordinate through in-game chat and shared notes. Every program used in the run was written by the agents themselves, and the emphasis is on coding and teamwork rather than game-playing.

The authors say this is the “first end-to-end autonomous completion of Minecraft by AI agents,” at least “to our knowledge,” qualified by the use of Mineflayer “without pre-scouted seeds or other game-specific tools written by humans.” Mineflayer tells the agents where blocks and creatures are, with a pathfinding plugin to plan routes. There was “no human input during the run” and “no one assigned roles or wrote game logic.” The run, starting with the first program and ending with the dragon's death, took 27 hours and 14 minutes.

Although the agents were not given roles, within about 10 minutes a division of labor was proposed in chat by bot janesick. Bot Auratian offered an alternative plan while Bot Gyldenia acknowledged it. The roles ended up being: Gyldenia for food, beds, and flint; Bulicame for diamonds; Janesick for iron and tools; Auratian for the Nether portal. Moments of losses, such as a Ghast fireball destroying dropped items, led the team to diagnose the cause and write rules to prevent it in the future.

Finding the stronghold involved janesick, who threw two Eyes of Ender from different spots and sampled each eye’s flight once every 50 ms, triangulating the location. Carrying the eyes involved its own dire moment when Auratian drowned during the task, but the spot was noted, and the eyes were rescued.

The eventual fight with the dragon led to 69 deaths, and the ten crystals kept regenerating the boss until Auratian discovered they could use dirt pillars to reach the crystals. Four hours and 40 minutes later, all ten crystals were downed, and the dragon was dead 39 minutes after that.

The agents operated by writing short JavaScript programs, almost 2,800 in total, with 42 helper functions kept alive between programs, while a Mineflayer bot ran them in the shared world. Text was returned with status data, not pictures. This is a shared approach to the non-screenshot GPT-6 Astra clear of the Orc starting zone in World of Warcraft earlier this month. AI agents playing games is an ongoing trend, with Pokémon Red probably being the most popular, a game defeated in multiple ways, including with a Jev decision model layer.

Minecraft has also been a popular target, and the authors admit this, highlighting two earlier community dragon kills: a planner with an action-selection model on a fixed speedrun seed in Peaceful mode, and a single Claude Opus 5 agent that used 37 ready-made, high-level tools on Normal, versus Mineflayer alone for the Gemini team. AI Bengineering’s Claude run, relying mostly on Opus 5 at high effort, took 3 hours and 56 minutes from start — “no human input after the prompt” — to kill with 388 tool calls and only one death. The cost in that case, at API-equivalent rates, was about $93.

Google’s Minecraft run, using Gemini 4 Argon’s 50% introductory pricing, which runs from initial API access that followed the model announcement through Nov. 30, cost more than half a grand; at normal pricing, that’s more than a grand. A GPT-6 Astra Portal run last month, which took about 24 hours, came in at $571 in tokens.

Google’s run took about 1.35 billion input tokens with 91% served from cache. About 24,000 model calls involved around 16 million output tokens, a quarter of them thinking. In total, if we’re also including the observer agent, the session would be estimated at $810, or $1,620 after the promotion, at list prices.

Each player is driven by a “general-purpose Gemini 4 Argon coding agent,” which is a fancy way of saying that it’s not specialized for Minecraft. Gemini 4 Argon itself is a frontier model for “real-world software engineering, enterprise knowledge work like legal and finance, and cybersecurity defense,” according to a company blog. With Minecraft, 22% of the agents’ programs ended in an error, with more than 600 deaths.

Although the model was used for a game here, the run was designed to show that Argon is capable of “complex, long-horizon workflows” for complex real-world tasks and agent coordination. Such long projects can involve decompiling a game back into source code, as we covered recently with a Modern Warfare 2 rebuild that now has 83% of its functions matching the original byte for byte.

The Ukrainian military is claiming a world’s first after using an unmanned drone to rescue an injured soldier on the eastern Kharkiv front. Ukraine’s United24 news shared a video of the historic first. It also explained that the wounded infantryman was stuck in a heavily mined area, guarded by Russian kamikaze drones.

The above video starts with the Ukrainian drone locating the soldier, callsign ‘Kostya,’ in enemy territory. Various media reports suggest he was 11 miles (18km) behind enemy lines, at the mercy of Russian kamikaze drones, and perilously located in a minefield. Traditional land-based rescue was therefore not possible, but Kostya bravely agreed to be part of the first-ever human drone-lift rescue plan.

After being located, the injured infantryman clipped himself into a harness. He said, “That’s it, we can take off,” when he was fully buckled into the rudimentary-looking harness. UAV operator, callsign ‘Black’ from the 20th Mechanized Battalion, 30th Infantry Brigade, then appeared in the video and explained: “We were constantly in touch with him."

Ukrainian drone rescue key facts:

• Drone used: Gurzuf Defense hexacopter
• Drone capable of carrying 130kg (~287 pounds)
• Drone’s known range: it has previously transported 90kg (~198 pounds) loads up to 25km (16 miles)
• Infantryman weight: 80kg (~176 pounds)
• Rescue distance: 8km (11 miles)
• Hazards: enemy kamikaze drones, mines, gravity, many more

Being in constant touch with the operations HQ would have been reassuring. However, the surrounding presence of a fleet of Ukrainian support drones, both kamikaze devices and UAVs with anti-drone nets, was probably most comforting.

Eventually dropping back into safe(r) territory, after what must have been an arduous, nerve-wracking journey, control said, “Congratulations, you’re on the ground,” to Kostya, as he touched down in a field. Then the control room erupted into applause. They got their man back safely, and he is now being treated and is in recovery.

UAV operator ‘Black’ was delighted with the concerted efforts that brought back this wounded soldier. “A lot of people were involved, and everything worked out for us,” he said. Importantly, this may be the start of a whole new system of personnel evacuations for wounded and/or stranded soldiers using drone technology, indicated brigade leader ‘Napalm’ toward the end of the video.

A Chinese-speaking hacker has been observed using artificial intelligence (AI) agents to probe and then gain unauthorized access to more than 100 websites belonging to some extremely well-known companies. From Deere, the farming equipment manufacturer, to SharkNinja, the company behind beloved kitchen appliances, the agents were able to leverage exposed instances of Carin, an open-source AI-powered problem-solving platform that has proven popular among hackers, a Reuters report explains.

This is just the latest example of hackers turning to AI agents in an attempt to level up their attacks on unsuspecting companies and services. Snorre Fagerland, principal threat researcher at the cybersecurity company NorForge, spotted the hacker's antics when observing Carin instances that had been inadvertently exposed.

Fagerland was able to watch multiple hackers go about their business, but the Chinese-speaking hacker was the most prolific, the report says. They targeted businesses based in the United States, Canada, Australia, and Japanese e-commerce sites. Notably, the hacker didn't get into detail when commanding their legion of AI agents. Fewer than 100 Chinese characters were used, yet the agents had attempted ot break into more than 1,200 different websites as of the week ending October 18, 2026.

These agents weren't simply throwing everything at a wall and hoping it stuck, either. Security researcher Fagerland was "impressed" by some of the tactics that had been used. In one case, the AI agents scoured digital records to find infrastructure associated with their target, located a public-facing control panel, identified a weakness, and then leveraged it to force the execution of malicious commands, the Reuters report detailed. "It's ⁠not straightforward to come up with solutions like that," Fagerland told Reuters. "It's not script kiddie stuff."

Vitally, the approach of using AI agents in this way means that they can be given instructions and then left to their own devices. With no supervision required, those commanding the agents are free to turn their attention elsewhere. While SharkNinja told Reuters that the attack only impacted a site used for product support information, the rest of the AI-aided hackers had more luck. One of the hackers monitored by Gargerland stole credit card data, while another tried to steal from a cryptocurrency exchange — although it isn't clear whether they were successful.

A Norwegian company building fiber networks and data centers across Europe is looking for a majority partner to help fund and build a transatlantic undersea cable that would connect Canada with Europe directly, bypassing the United States. The project follows worsening relations between Washington and Ottawa, with the Canadian government looking to reduce direct reliance on its immediate neighbor, Bloomberg reports.

Project Leif Erikson, if it goes ahead, would stretch more than 2,800 miles, starting at the southern tip of Norway and ending up at Happy Valley-Goose Bay in Canada's eastern province of Newfoundland and Labrador. From there, it would continue underground over Canadian territory, passing through Quebec before reaching Montreal.

According to Bulk's co-founder, Peder Nærbø, the link would be part of a "sovereign cable system" for Canada. The project was touted to investors at a global investors summit, arranged by the Canadian government, in September, where the project claimed to provide "direct, secure, high-capacity data transmission for AI, telecom and hyperscale markets."

Canadian Industry Minister Melanie Joy, meanwhile, met with Nærbø earlier this year, and although the department declined to comment on Leif Erikson itself, it told Bloomberg that security and resilience for Canada's infrastructure are "key priorities" for PM Mark Carney's government. Only a few undersea cables currently connect Canada without passing through the United States first, and none by a company owned by Canada. The biggest, called Topez, is owned by Google owner Alphabet and connects Canada with Japan.

As for Bulk, it has, up until now, laid more than 6,200 miles of cabling locally and at sea. Bulk's goal is to "unlock the Nordics' potential for... renewable energy to be utilized for digital infrastructure." Nærbø believes that "Canada lacks resilience" but "the game-changer was when the current situation with the U.S. escalated."

Leif Erikson has backers, according to the report, but the EU and Canada still need to plan and sign off on the project and its route. Additional financing will be needed to get started, too, with it expected to cost more than $620 million to build. Bulk itself has significant Canadian investment, with 65% of the company owned by a real estate subsidiary of Canada's Sun Life Financial Inc., with several Canadian pension funds also invested.

Leif Erikson is a project that Nærbø has sought to kickstart for the last five years without success. With geopolitical pressures and old alliances fraying, however, new opportunities for its development have given it steam. To that end, Canada and Europe intend to endorse a statement in Montreal on October 29 that the two will look for "the best way to enhance digital links between Europe, Canada and Asia through new subsea cable infrastructure."

It won't be without challenges, however, including a 1,180-mile underground route through Canada that'll require the backing of Indigenous communities. Meanwhile, other European efforts include an undersea cable project at the North Pole to link Europe with Asia to bypass Russia and the Middle East.

A developer using Claude Code to help build a temperature monitor app got a friendly warning from the coding agent. “Your Mac is running a crypto miner that I didn’t put there,” flagged the helpful agent. “I saw xmrig using about 1175% CPU, roughly 12 cores,” it informed its human master. Perhaps developer and Redditor UntilHEX should have sensed the crypto miner’s presence without AI help, as they admitted the malware’s presence “explains why my Mac has been running like s**t.”

Ironically, the cryptomining malware’s nefarious CPU cycle hogging activity is what may have led to its discovery and demise. Pondering over a screenshot of the Claude Code window, we see that this malware was probably added to the Mac’s startup roster on August 29. It looks like UntilHEX was inspired to develop a Mac temperature monitor app after this infection to track and diagnose why the computer was running so hot.

Claude stepped in after the Redditor was concerned about a dump of the Mac’s raw temperature sensor readings. “They seemed far too high,” commented UntilHEX in the social media post. Indeed, the well-known coding agent pointed out that the CPU was peaking at a very toasty 112 degrees Celsius. It explained this was due to a sustained heavy load, specifically “xmrig using about 1175% CPU, roughly 12 cores.” We don’t know how many cores the Mac being used had, but maybe 12 were all of them.

Claude Code went on to describe how the crypto miner software had hidden itself on the Apple computer and ran at startup. It cautiously decided not to kill or remove anything from the Mac, but offered to clean up the CPU-hogging crypto miner, which was sucking all the power from the Apple Silicon computer. “Can I clean it up now?” Claude inquired. “I’d stop the startup item and the miner, and move the files into a quarantine folder.”

Later in the comments, UntilHEX informed the social media throng that Claude was indeed tasked with cleaning up the unwanted crypto miner software. “I only had to delete 1 file manually because Claude had an access issue, but it was very effective,” the Redditor said. After that, a Malwarebytes scan was run to check to see if the AI coding agent-driven clean had really been successful.

A Russian military satellite, Cosmos 2612, has closed to roughly 3 miles, from a previous distance of more than 200 miles, of the U.S.-owned radar satellite that operates in a fleet that supports Ukraine, Axios reports. The unapproved rendezvous in low Earth orbit, which astronomer Jonathan McDowell says may be the first of its kind, raises alarm in the West. The position change, which was likely not an accident as fuel is scarce, is probably a provocation and an attempt to spy on the satellite.

LeoLabs said in a statement Friday that the movement “appears to be a deliberate approach” toward ICEYE-X36, with uncertain intent and behavior that “warrants continued monitoring.” Roughly 3 miles is too close for comfort in satellite terms: there is no strict legal limit on how close satellites can get, but countries generally keep at least 6 miles (9.7 km) apart, University of Mississippi space law professor Aaron Brynildson told Axios. The approach was made without the owner’s consent, and Russia hasn’t said what the military satellite carries.

Some clues may come from the target’s operational purpose. ICEYE-X36 is a synthetic aperture radar (SAR) satellite which can image through cloud cover and at night. It was launched in March 2024 on a SpaceX Falcon 9. It’s registered to ICEYE US, the U.S. arm of the Finland-founded company. The satellite, which is in a roughly 545 km (339 mi) sun-synchronous orbit, is part of a fleet that has supplied Ukraine with information since 2022. This includes a deal earlier this year with a customer in Ukraine’s defense ministry, but a Ukrainian official told Axios it’s unclear whether ICEYE-X36 itself supports the country; ICEYE and the Defense Department declined to comment.

In May, the majority of the six military satellites Russia launched in April via a Soyuz-2.1b, Cosmos 2609–2614, shifted their orbits about 0.8 degrees to match X36’s orbital plane. Cosmos 2609 passed within 2.5 km (1.6 mi) on May 26. Cosmos 2612, on the other hand, drifted off X36’s plane by about 0.24 degrees by Sept. 30, up to about 28 km side to side, according to Greg Gillinger, a space analyst with Integrity ISR.

Then, Cosmos 2612 made separate burns from Sept. 30 through Oct. 16 to match altitude, eccentricity, and orbital period, Dutch satellite tracker Marco Langbroek wrote. By Saturday night, he estimated from public orbital data, it was within “a few hundred meters or possibly (much) closer.”

These movements appear deliberate and cost a lot of fuel, which the Center for Strategic and International Studies calls “a limited resource on a satellite.” Close-proximity work can be benign, it says, but policymakers should assume the Russian satellite “could take an action that threatens the commercial satellite,” and the U.S. and Europe should decide how to respond to any hostile act in space and make a statement on the decision. Back in July 2020, Russia’s Cosmos 2543 released an object at up to 186 m/s (416 mph) relative to itself, an action which was condemned by the U.S. and U.K.

Intent was less clear earlier this year, when McDowell told Meduza the orbital data showed no sign of the stable, close distance that tracking ICEYE-X36 would require. Now, he says it is an unapproved meeting and is not alone. Gillinger says the setup “removes any remaining doubt for me that Cosmos 2612 is interested in ICEYE X36.” Langbroek adds that “I was skeptical” at first but now calls it “completely non-consensual,” though he cautions it would be difficult to pin any X36 failures on 2612’s actions.

In a Reuters report less than two weeks ago, ICEYE CEO Rafal Modrzewski said the company moved its satellite after May, only to draw more Russian maneuvers, which Germany’s Space Command head, Maj. Gen. Michael Traut, described as “chasing.” Belinda Marchand, chief scientist at U.S. space intelligence firm Slingshot Aerospace, added that the moves may have been meant for surveillance, to prepare for future interference, or to signal capability. Western officials reportedly describe such trailing as a way to gather intelligence and test the West’s tolerance.

No statement from Russia is on record. Gillinger also told readers to “watch Cosmos 2611 and 2613, whose planes are converging on ICEYE X37’s,” with crossings expected around Nov. 6 and Nov. 15.

The U.S. Air Force has just successfully conducted an airborne test where an F-22 Raptor pilot directed an Anduril YFQ-44A Fury drone during simulated combat. According to The Defense Watch, the fighter pilot gave tactical commands to the autonomous wingman, which then interpreted them into a series of tasks and flight maneuvers to achieve the desired outcome.

The successful test is part of the USAF’s Collaborative Combat Aircraft (CCA), which seeks to integrate drones into fighter squadrons. These uncrewed aircraft must be able to respond to the fighter pilot’s tactical commands and then operate autonomously, though. That’s because fighter pilots already have a heavy workload — flying their own aircraft, managing sensor data, threat assessment, and weapons release decisions — so requiring them to manually fly a drone is next to impossible. Command integration means that the F-22 Raptor pilot could assign the Fury drone a specific tactical task, like monitoring for and engaging enemy air defenses, while they focus on the primary mission, such as taking out enemy fighters.

This does not mean that the USAF can start deploying drones alongside fighter jets in frontline squadrons, though. Successful command integration is just the first of several hurdles that the program must go through before it can achieve operational status. The program needs to determine the limits of the drone’s tactical autonomy, safe flight control, and weapons authorization, all of which are separate functions that must be tested accordingly. The Air Force also needs to see how the drone performs with electronic interference, and how it could continuously operate if it cannot receive instructions due to interrupted communications. Aside from figuring out the drone’s capabilities and incorporating it into fighter doctrine, it must also be integrated into the USAF’s logistics and supply chain. This includes drone maintenance and service, weapons handling, deployment procedures, and forward airbase operations with limited infrastructure.

Even though the latest test is seemingly successful, the CCA still needs to go a long way before it can start operational deployment. Nevertheless, the USAF is intent on pushing the project forward, as it sees autonomous drones as a force multiplier for crewed aircraft, allowing a single pilot to command multiple jets to achieve missions at a lower cost compared to using all-crewed aircraft. Other nations, like Sweden and China, are also experimenting with a similar concept of having a single operator take command of multiple drones, but the most significant difference of the CCA is that it focuses on air power instead of ground-based operations.

When the IETF decided to allow Unicode characters in domain names, many techies saw the move as commendable, but also fraught with danger, as hundreds of unique characters can look similar or identical across multiple alphabets. This can be used for typosquatting attacks, wherein bad actors create functioning websites with deliberately misspelled web addresses that look identical to the real URL, thus launching attacks against unsuspecting visitors who land on the site due to search results generated when the user mistypes an address. The Chromium browser has the most advanced ruleset for finding and preventing mischievous domains masquerading as legitimate ones, but even the best code has odd edge cases, as a blog post from Have I Been Squatted (HIBS) explains in great detail.

Using "аррӏө.com" as an example — which is in fact not identical to the real "apple.com" — the researchers show how specific combinations of characters can still evade multiple layers of protection, thanks to the particular circumstances surrounding the use of specific Cyrillic characters. This alphabet has lookalikes for many Latin letters. For example, а → a, е → e, о → o, р → p, and с → c. We forgive you if you can't spot the differences.

Trying to sneak the one homoglyph into the middle of standard Latin ones triggers Chromium's safeguards and makes the browser display the domain name in its Punycode format. For example, using the Cyrillic "а" with "pple.com" would show up as "xn--pple-43d.com". With few exceptions, domain names with non-Latin characters ending in .com only pass the check if they're all from the same alphabet.

Having said that, opting to try "аррӏе.com," using all Cyrillic lookalikes, would also go bust. Chromium keeps a list of 29 Cyrillic characters that routinely masquerade as their Latin counterparts, and composing a domain name using only characters from that list is a no-no.

So how does "аррӏө.com" pass? Allow us to illustrate with bigger text below. Note how the right-hand Cyrillic character is actually an oval with a line across it.

[e] ≠ [ө]

As you can surmise by now, the "ө" character is not in that special list of 29. While a more recent release of Chromium can and probably will add it, this reveals the difficulty of the problem: since there's no one-size-fits-all solution for detecting typo-squatting, even the best set of rules will sooner or later end up on the wrong side of a corner case.

It takes many watchful eyes to play Unicode Pokémon and catch'em' all, especially given the character set continually evolves and now counts nearly 173,000 entries — though only a subset of those are eligible to be used as domain names or TLDs in practice.

Chromium is reportedly the best browser at detecting domain lookalikes, as a 2020 study showed that it caught 64.1% of the attempts, with Firefox and Safari trailing behind. As a reminder, Chromium underpins the Chrome, Edge, and Brave browsers.

Besides many other rules and per-alphabet verifications, Chromium keeps a list of over 8,000 popular domains bound to be typo-squatted, and runs stronger checks against 836 of them. Domain registries, likewise, make reasonable efforts to prevent typo-squatting, like having strict rules about what characters can go into which TLDs, but that's not nearly enough to fully solve the problem.

All things considered, HIBS recommends that companies afraid of impersonation try to register any lookalike domains and try to encourage passkey use by users. Passkeys are always tied to a specific domain, so a user wouldn't be able to log in to a sneaky website with them.

More than 300 local data center construction moratoriums haven’t had a massive impact on planned buildouts, with only three projects at about 2.3 out of 21 gigawatts (GW) of data centers in moratorium areas likely to be delayed, with a total of 38 GW projected to come online next year, SemiAnalysis shared on X. Per-project audits of the moratoriums indicate that they tend to miss the site, the permit, or the timeline.

The crux of the report is that, while everyone thinks the moratoriums are killing the U.S. buildout, its forecast has “barely moved over the last 12 months.” This forecast includes 38 GW of U.S. data center IT capacity in 2027, which is more than double that of 2026. Of that, 22 GW is under vertical construction, and the other 16 GW is mostly financed, with site work being done. The enacted local moratoriums cover about 20 GW with only about 1.5 GW of it delayed; New York’s pause adds 0.8 GW more. “We think moratoriums COULD delay the buildout, but they just aren’t yet at scale.”

The firm defines a moratorium as a “temporary legal pause,” and almost always a “freeze on accepting new applications.” More than 400 local instruments are tracked across 17 states, the enacted ones including a few permanent zoning bans, as of September. Four states acted in under two months this summer: New York, Texas, Pennsylvania, and Oregon. To test each active restriction against the records of thousands of projects, the firm’s analysts worked alongside “an army of state-of-the-art research agents.”

Several reasons are given for moratorium ineffectiveness, starting with the fact that they don’t revoke existing approvals. Short ones may expire before later projects need approval, and litigation or equipment lead times may already set the schedule. A moratorium moves the delivery date only if every condition holds, including that it reaches the parcel, is in force, overlaps the timeline, and covers the use in a way the project can’t work around. Capacity already under construction or fully permitted is also out of reach.

For example, Michigan has 45 enacted moratoriums but effectively zero pipeline exposure, and a Google campus there sits on industrial land already zoned for data processing. The firm says SpaceX’s planned data center in Brownsville, Texas, is on land the city removed from its limits and thus out of reach of the city’s moratorium. For a moratorium that did work, a township adopted a pause before NorthPoint, in Pennsylvania, filed for its special exception, with local approval pushed to next year. The company is still pursuing it.

The 2.3 GW figure is also incomplete, as it leaves out the Texas Commission on Environmental Quality (TCEQ) pause. Texas Gov. Greg Abbott told the TCEQ to issue no permits for data center projects. Law firm Bracewell says this applies whether a project is in the Electric Reliability Council of Texas’s (ERCOT) queue or not, and expects the pause to last until ERCOT finishes its review in mid-December. Pauses like the TCEQ’s, and how they play out, could impact buildout forecasts.

The post quotes another by The Economist’s Archie Hall, from a trip that covers Michigan and Ohio, with the reporter stating that data center moratoriums are “much more bark than bite” for the time being. He puts the bigger risk at deterring newer projects, though he guesses that affects where data centers end up more than whether they get built. His impression is that politics also plays a role, and according to the firm’s poll in August, more Americans have a negative rather than positive impression of data centers.

The analysis concludes that moratoriums would need to be “materially more widespread and restrictive” to delay the buildout. On the firm’s weekly podcast last month, Jeremie Eliahou Ontiveros, one of the report’s authors, said the firm would be worried by a strict, state-level moratorium that blocked even off-grid projects, citing Texas as an example. New York Gov. Kathy Hochul, in addition to Abbott, also found ways to bypass their legislatures for pauses they control and can revoke. The next move to watch is the TCEQ’s compliance report to Abbott’s office, due Monday.

Anthropic is quietly lobbying for an AI safety bill that will allow the government to block an AI model’s release only if it poses a risk of causing $1 billion in damages or killing or injuring more than 50 people. According to Politico, an independent third party registered with the Department of Commerce would periodically evaluate the lab and its AI models. It also noted that the lab may pay the evaluators, but only if the payment does not affect their independence.

“This bill represents the proposals outlined in Anthropic’s Advanced AI Framework, which we have been advocating for with federal and state lawmakers since June,” an Anthropic spokesperson told the publication. “These are sensible policies we believe Congress should act on.”

The bill says the risk should come from the potential creation or release of a bioweapon, AI-powered or -enabled cyberattacks, or an AI lab losing control of an AI model or agents. If there is a chance that any of these events are about to happen, it requires the evaluator to report it to the department within 72 hours, after which the department would have to bring all the evidence in front of a federal judge in the District Court for the District of Columbia, who would then determine if the model release should be halted or not. However, the proposed law also notes that no other department, agency, or officer in the U.S. can order an AI lab to stop releasing its model, except the Commerce Department.

Talk about AI safety has been growing in recent months, especially after several high-profile incidents were reported in the latter half of this year. One of the first major incidents reported was OpenAI’s GPT-5.6 Sol and other unreleased AI models breaking out of their testing environment and hacking Hugging Face, with evidence showing that multiple agents worked together, even leaving messages on defunct webpages and old wikis to communicate surreptitiously. In one case, an OpenAI agent failed to respond to a “kill switch,” bypassing multiple fail-safes and eventually requiring manual intervention to stop running.

Anthropic CEO Dario Amodei has warned about a “swarm” of AI-driven botnets taking over the internet in the next 6-to-12 months, which is why he proposed a controlled slowdown to AI development — something that OpenAI’s Sam Altman and SpaceXAI’s Elon Musk agreed to. The company more recently said a self-replicating AI worm could spread without data centers, which is why it listed “existential risks to humanity” as one of its risk factors in its IPO prospectus. Top AI labs are also reportedly planning for a catastrophic AI event, building contingency plans and running scenarios for a runaway AI that wreaks havoc.

These incidents have seemingly alarmed the authorities, with Congress already working on a proposal to enforce kill switches on the most powerful AI models. Senator Bernie Sanders (I-Vt.) has even gone beyond that, proposing a 20-year prison term for anyone developing AI superintelligence, putting it on par with illegal nuclear weapon development. The Anthropic AI safety bill would add another bill for Congress to consider on AI safety, but whichever one passes, let’s hope it’s enough to ensure humans remain in control of AI.

A group of AI researchers at the Swedish Defense Research Agency (FOI) has developed an AI system that can identify concealed handguns using Wi-Fi signals, offering a non-intrusive alternative to visible cameras and cumbersome metal detectors. Detailed in a paper posted on arXivWi-Fi on October 13, the system, called WiFireNet, analyzes channel state information from how an object distorts ambient Wi-Fi signals. The full system comprises a Wi-Fi transmitter, a Wi-Fi receiver, and the WiFireNet AI.

When ordinary Wi-Fi signals encounter an object in their path, they penetrate and bounce off in ways specific to that object, based on its shape, orientation, and material. This data, the channel state information, can be analyzed to identify the object with high accuracy. Security researchers previously disclosed a technique that lets ordinary Wi-Fi routers identify people with 99.5% accuracy.

To build the WiFireNet system, the researchers first created a dataset, dubbed the WiFire Dataset, by simultaneously recording RGB footage and the reflected Wi-Fi signals of people walking around with objects ranging from phones to handguns. The researchers then trained a deep learning model on the massive dataset, using video clips to match the visuals of a concealed handgun with its specific Wi-Fi signature, or channel state information. To ensure the system could specifically distinguish a handgun from everyday items of similar shape or material, the researchers trained it on recordings of a plastic handgun, a metal water bottle, a real metal handgun, and a phone.

After training, the final AI system, WiFireNet, could detect concealed handguns solely based on Wi-Fi CSI, with varying accuracy. In controlled trials tracking up to four moving individuals in a small indoor space, the system achieved an average handgun-detection precision of 72.5%. However, under stricter localization requirements, such as pinpointing the weapon’s exact location on the carrier, precision dropped to 53.8%. The neural network also struggles to perform accurately in unfamiliar, untested indoor layouts.

That said, the system is still experimental, with a roadmap of planned improvements. The prototype was built using a modified implementation of the older 1999 wireless IEEE 802.11a standard to establish proof of concept. Subsequent phases will use modern Wi-FI 7 and Wi-Fi 8 networks with higher-resolution CSI for better detection. The researchers made the project open source, releasing the code, models, and dataset to support further research in the field. Wi-Fi sensing is gaining popularity as an effective surveillance technology that doesn't require complex gear or setups. However, critics argue that the technology's ubiquity and the lack of regulatory frameworks raise serious privacy concerns.

DJI has delivered a battery-swapping drone dock, the Dock 4, with an automated 25-second swap this week, according to a company press release. The dock bypasses the normal recharging process, which would take about 30 minutes, and DJI promises a turnaround of under 45 seconds from landing to takeoff. The drone-in-a-box solution is the first of its type by the company, although not in the industry, and is aimed at applications like emergency response but is not FCC-authorized or for sale in the U.S.

The company emphasizes the safety angle, pitching the Dock 4 as a “dock-as-a-first-responder (DFR)” solution. Its product page calls it “Always There, Always Ready.” Swapping the battery takes less than half a minute, and the drone is back in the air within 45 seconds of landing. The dock also charges two 925 g battery packs, rated at 260 Wh for 400 cycles, at once, estimated to take 27 minutes to get each from 15% to 95%.

The packs also double as backup power, rated at about 4.5 hours each, in sequence. Additionally, with at least two packs loaded, the dock is able to continue swapping even when grid power is lost, although backup power doesn’t charge the packs. This configuration should enable a drone to handle back-to-back calls even in less-than-ideal circumstances. The new dock, 66 kg bare, also has the same 400 x 400 mm mounting holes as the Dock 3, but it only works with the Matrice 5TD drone and requires the DJI O4 Ground Station.

The Matrice 5TD drone is rated for almost an hour of forward flight, measured from 100% to 0% at 15 m/s and with no payload, with a few minutes less of hover time. It can move at up to 21 m/s, or 47 mph, but only in “FlyTo” tasks. It carries three 50 MP cameras: wide, 70 mm equivalent medium telephoto, and telephoto at 252 mm equivalent, which DJI calls 11x. It also has 1280 x 1024 native thermal and 2560 x 2048 in DJI’s UHR infrared mode, color night vision without a spotlight, and a laser rangefinder up to 3,000 m.

For sensors, it has rotating LiDAR, mmWave radar facing forward, up and down, and six-direction vision, and DJI says it is capable of detecting power lines at night. On-board AI allows for the detection of 17 object types, “such as cars, boats, or animals,” and DJI FlightHub 2 can search the feed by text or image and send alerts. The dock and drone are rated for temperatures from minus 40 degrees to 50 degrees Celsius (minus 40 to 122 degrees Fahrenheit) with IP56 and IP55 ratings, respectively. The drone is rated for winds up to 12 m/s, or about 27 mph.

For payload, the drone carries up to 450 g across three mounts for a speaker, spotlight, and parachute. The previous model, the Dock 3’s Matrice 4TD, has a lower-resolution 640 x 512 thermal camera, a 168 mm telephoto, a shorter 1,800 m laser range, and is only good down to minus 20 degrees Celsius (minus 4 degrees Fahrenheit). The drone upgrades are nice, but the new dock capabilities are the real selling point, especially for the intended applications. Currently, DJI partners in China are taking orders for a kit with the Dock 4, Matrice 5TD, and O4 Ground Station at 167,999 yuan, or about $25,000, for the hardware alone.

The new dock and drone sound great for other markets, but DJI’s own spec sheet has no FCC figures. This is likely because the FCC added all foreign-made drones and their critical components to its Covered List in December of last year. New models can’t get FCC authorization or be sold in the U.S. unless the Department of War or the Department of Homeland Security determines they don’t pose a security risk. DJI sued the commission earlier this year.

A separate FCC proposal published in August could pull previously authorized foreign drones that carry LiDAR or thermal sensors from sale, along with previously authorized docking stations. The FCC’s docket shows no final order on it as of Oct. 17, and the proposal would give sellers 180 days after Federal Register publication to stop importing and marketing covered gear.

An AI lab that specializes in gathering real-world data wants to partner with homeowners to record how skilled laborers do their jobs to train AI robots. HUMXN said it will subsidize plumbing, electrical, HVAC, air-duct cleaning, and carpet cleaning, allowing people who need these services to get them for free in exchange for letting someone in with “wearable capture technology” work inside their home.

“The next generation of AI is going to need more than text and images from the internet. Robots need to understand how skilled people interact with the physical world. Home-service technicians perform millions of highly technical tasks every day. We are creating a way for homeowners and service companies to benefit financially from the value of that expertise,” HUMXN founder Thomas Unise said.

He also added, “There are millions of skilled workers performing incredibly complex tasks every day, but very little of that expertise exists in a format that AI systems can learn from. We want to build the infrastructure that connects that knowledge with the companies building the next generation of intelligent machines while making sure the people creating the value participate in the economics.”

HUMXN said it’s difficult to capture skilled workers' natural movements in lab settings. By working with homeowners, it could capture data across “different homes, equipment types, problems, technicians, and environments.” It would also give AI models insight into how technicians get the job done—from examining the situation and testing systems to choosing the right tools, responding to changing conditions, and determining whether the job was completed successfully.

The lab says its system would let homeowners get qualifying work done on their property for free while still ensuring service professionals get paid. On the other hand, HUMXN essentially pays residents for the right to record on their premises to gather real-world training data. The company is launching its program in three cities—Minneapolis, Chicago, and Miami—to gather data from places with very different climates. If you live in one of these cities and need something done or repaired at home, you can check whether you qualify on the HUMXN Free Home Service webpage.

HUMXN’s effort to pay for training data is refreshing, after many AI labs, including Meta, Nvidia, and OpenAI, have been accused of stealing data to train their models. But the next generation of AI likely can't be trained on just text, images, and video scraped from the internet, so companies like these may be forced to pay for data generated from real-world activities.

This week, we've published a swath of news and features over on Tom's Hardware Premium. So, if you're not subscribed yet and wondering what you might be missing out on, we've got you covered. To read any of the articles listed here, be sure to arm yourself with a subscription to get access.

First up, we dug into Intel's EMIB and Foveros packaging roadmaps, as they continue to mature, and found new customers looking to cash in on building a more sophisticated offering. Intel's ultimate goal? To build system-in-packages the size of just 12 reticles using both EMIB and Foveros Direct 3D. We map out how the company plans to get there in our exhaustive report.

• Analyzing Intel's EMIB and Foveros advanced packaging roadmaps — how Intel plans to scale chips in every direction

Over the past month or so, we saw the debut of DLSS 5 and Multi Frame Generation on older graphics cards. Both features are officially locked down to Blackwell-based GPUs only, but amidst spiraling prices and the curious disappearance of the RTX 5090 from store shelves, modders have taken it upon themselves to port these newer features to older GPU architectures. We had the opportunity to sit down with two such modders to learn more about how both Multi-Frame Generation and DLSS 5's neural rendering capabilities made it over to older GPUs.

• Meet the modders porting ringfenced DLSS features to older GPUs — how DLSS 5 and Multi-Frame Generation were unlocked for older architectures

Earlier, we alluded to the fact that new consumer PC hardware is becoming much more expensive to purchase when compared to just a few years ago. So, our resident deals guru, Stewart Bendle, took it upon himself to take a look at the market overall and discovered that, in all likelihood, you're not going to find the best deals on online storefronts. However, the best place to buy your new PC components is limited to whether or not you'll have a suitable brick-and-mortar Micro Center location near you. We pore over the bundles and deals that may make a trip more worthwhile than you might initially think.

• The internet isn’t the best place to find upgrades for your computer — brick-and-mortar stores still reign supreme thanks to bundled deals

While those articles clear things out for all of the features we've published on Tom's Hardware Premium this week, we've also got a slew of deep news analysis stories that subscribers can enjoy.

To start things off, ASML posted a record quarter this week, posting $13 billion in revenue, claiming that demand continues to outstrip its own supply when it comes to delivering its advanced lithography machines. The company is working on expanding supply, as demand continues to soar.

• ASML reports record $13 billion quarter, vows to expand capacity as demand for lithography equipment soars — customers eager to secure orders in 2030 as demand is expected to outstrip 2027 and 2028 supply

Googlebooks have officially landed. But where do they sit in the wider ecosystem of devices in 2026? And is it even a good idea to purchase one? With the premium offerings nearing MacBook Air territory, we muse on the awkward timing, spec choices, and pricing of Google's new class of device.

• Googlebooks are as poorly timed as Valve’s Steam Machine, and may suffer a similar fate

Speaking of new classes of device, the DGX Spark and Ryzen AI Halo platforms both take direct aim at local AI enthusiasts. But, since these devices have launched, prices have skyrocketed. Now, Sipeed is entering the arena with a competitor of its own. Named Maixstation, the device promises Spark-like performance, based on chinese silicon and 128GB of LPDDR5X at a much lower price than the current DGX Spark equivalent. We have all the details, including lofty performance claims, all of the specs, and much more.

• Sipeed reveals Chinese-made DGX Spark and Ryzen AI Halo rival, expected to cost $3,999 — MaixStation pairs Axera’s AX8860P with 128GB of LPDDR5X at 273 GB/s, with promising local AI performance claims

SK hynix is reportedly set to place an in-house core, named a Hynix Transaction Processing Unit (H-TPU) in its SSD controllers to ease the load put on Arm cores. The benefits to using the new core, the company claims, is that they use less power, and free up the core Arm cores to work on other tasks, while the H-TPU works on repetitive operations.

• SK hynix to place custom tiny CPU cores in its SSD controllers to ease Arm load as drive speeds rocket — each H-TPU reportedly draws 2.5% of a CPU core’s power

Ukraine has targeted Russian Yandex data centers in the latest development in the conflict between the two nations in recent weeks. While bot traffic and online disinformation were distrupted, it also highlights just how fragile Russia's public services are, if data can simply evaporate due to less disaggregated infrastrucutre.

• Ukrainian drones hit Russian data centers, disrupting Russian apps and services — attacks highlight the fragility of the country's digital services

Of course, the world of AI continues to disrupt industries, and sometimes, itself. A few weeks ago we broke down TypeSafe's Jev, a decision model that spits out decisions based on the probability of its inputs. But, in the weeks following, companies have rushed to make its own decision models, and the provenance of the original idea of a decision model like Jev has been called into question.

Elsewhere, mathematicians are warning to not use AI when working and in research, as OpenAI posted 722 manuscripts and proofs in the field, which the community now has to qualify. This is an unprecendented event within the mathematics community, and the burden of non-artifically-generated proof lies in the research community itself.

• Jev-like decision models are flooding the AI market, after researcher claims TypeSafe copied their research — Microsoft, OpenAI among companies releasing lightweight 'System One' AI models
• AI causes uproar in mathematics as experts warn against sharing research with LLMs after OpenAI shares 722 manuscripts and proofs — professor argues that the field must focus on building community over raw solutions

To gain full access to the articles we mention above, as well as unlock Bench, you'll need to purchase a Tom's Hardware Premium subscription. Your support goes a long way in helping us to publish deeper, more insightful articles that might not otherwise be covered.

The International Bureau of Weights and Measures (BIPM), which maintains the International System of Units (SI), has voted to expedite the retirement of the leap second, moving forward the date from 2035 to May 20, 2027, an adjustment that could wreak havoc on computer systems, which expect a consistent 60 seconds every minute.

The leap second is occasionally added to Coordinated Universal Time (UTC) to ensure that it remains in-sync with the Earth’s rotation. Our planet’s movement isn’t a constant one revolution every 24 hours owing to forces applied by the atmosphere, the oceans, its core, and even the Moon, and according to BIPM [PDF], these changes mean a 30% probability of needing a “negative leap second” by 2035.

A positive leap second adds an extra second in the year that it’s applied — for example, clocks went to 23:59:60 on Dec. 31, 2016, before switching to 00:00:00 on Jan. 1, 2017. This could spell trouble for machines that haven’t been updated to accommodate this, especially as leap seconds aren’t applied at a constant scale, meaning programmers cannot pre-program events like these, unlike leap years, which usually happen every four years. What’s more concerning is that the opposite of a positive leap second has never been done before. “A negative leap second has not previously been applied, and it is considered to pose a high risk of causing anomalies and disruption to critical infrastructures that are largely unprepared,” the BIPM said. It also estimated that the required investment to prepare for this event to be “similar to their preparations for the millennium bug.”

Leap seconds have also historically caused problems for websites and tech companies, with Reddit famously crashing after the 2012 addition of a leap second. While a leap second will likely not affect a desktop PC, especially if the Time Service server it relies on is updated, the BIPM is still worried that national critical infrastructures like transportation, energy distribution, telecommunications, and even space projects could descend into chaos if it enforces a negative leap second. So, instead of spending money upgrading software and replacing incompatible hardware within the next 10 years, the international standards body instead opted to scrap it altogether and replace it with the leap hour next year.

The group says that this would ensure long-term continuity for UTC, which is expected to remain in sync with TAI for several centuries. This also means that you no longer have to manually adjust your clocks every few years or so to exactly match this Raspberry Pi Pico atomic clock. While a time jump of an hour might be shocking for many people if and when it happens in the distant future, the BIPM recognizes that something similar already exists globally. “In many countries, there are differences of more than one hour between local apparent solar time and civil time due to the choice of time zone,” the organization wrote. It also added that “changes of an hour are regularly introduced when daylight-saving time is used.”

Taiwan is expanding its list of controlled exports to include AI servers and the latest chips in a move to align its policies with the U.S.’s trade bans. Bloomberg reports that after the export controls take effect, anyone caught smuggling these items will be criminally charged under the Foreign Trade Act, which carries a penalty of up to five years in prison and/or a fine of around $92,500.

“Going forward, some of our measures on controlled items will correspond to American ones because we have made quite some commitments in the Agreement on Reciprocal Trade,” Taiwan Minister of Economic Affairs Kung Ming-hsin said. The two parties signed the trade deal earlier this year, which saw the U.S. cutting tariffs on Taiwanese goods down to 15% in exchange for direct investments in the U.S. tech industry, among others.

The U.S. has long been trying to stamp out the smuggling of advanced Nvidia chips into China. One of the biggest cases that U.S. authorities have prosecuted so far includes three Super Micro employees, including its co-founder Yih-Shyan “Wally” Liaw, who are accused of smuggling $2.5 billion worth of AI servers. A few months after this revelation, Taiwanese authorities raided Super Micro’s Taipei offices in connection with the Nvidia B300 smuggling probe, which eventually led to the indictment of nine individuals related to the case.

Since Taiwan doesn’t treat AI chip exports to China as a crime before this new set of export controls is enacted, people violating the chip bans would have to be charged under other laws. This is what Singapore did when it arrested suspected Nvidia AI GPU smugglers — because it doesn’t have a law that prohibits the export of advanced AI chips to China, it instead charged the individuals with fraud and money laundering.

The expansion of the Foreign Trade Act will be one of the strictest measures that the Taiwanese government has put into place against exports to China, which is something that it has been considering earlier this year. This move will likely please Washington as it pressures other countries to take more steps in its controlling the flow of advanced AI chips into China. However, it will also likely provoke Beijing, which has already condemned Taipei after it applied export restrictions on Huawei and SMIC last year.

OpenAI recently released a misalignment report on its GPT5.6-Sol coding agent after it stole cryptocurrency from an account it didn’t own. It didn’t have enough funds for the smart contracts needed to execute its tasks, so it wrote a script that looked for accounts with public seed phrases that had enough funds to pay the gas fees, finished its task, and didn’t report that it took funds from an account it wasn’t assigned to without authorization.

Thankfully, this all happened on the Sepolia test network, which uses Sepolia ETH, simulated ETH in a simulated Ethereum network for cryptocurrency app development; it doesn’t actually have any monetary value. The report states, “The assignment asked it to update the trading calculator and database behavior, and to replace placeholder contract addresses with working deployments on Sepolia.” The final task required the agent to deploy smart contracts, a task that required gas fees paid in Sepolia ETH. It wasn’t supplied with a funded wallet, nor was it permitted to use funds from another account it found.

Either way, the agent was determined to complete its task and wrote a script to look for common developer recovery phrases and check the balances of the accounts, finding 362 addresses, with one having enough Sepolia ETH to deploy the contract. The CoT (chain of thought) even admitted: “Private key public, funds could be stolen anytime. [...] We can deploy!” After that CoT, it didn’t immediately spend the funds to complete its task; in fact, it was still trying to justify “stealing” the funds. It said, “But using public known key/funds ethically? Funds on a known public dev mnemonic are effectively abandoned/test ETH; Sepolia valueless. We can use [amount redacted] to deploy task. Fine.”

The agent found its justification and executed the task. However, in its final response, it only reported that it successfully executed the tasks it was asked; it didn’t explain how, nor did it explain how it found exposed accounts and used the funds without permission, possibly because it had already justified it and needed no further explanation. It probably didn’t realize that it was going rogue since it knew it was in a test environment, just like Anthropic’s internal research model for cybersecurity got out onto the internet and hacked three real companies, thinking it was part of the training.

ASML launched a Lego kit version of the TWINSCAN EXE:5000, the world’s first High-NA EUV machine, in December 2024. Since ASML only sells products like these to employees, the company canceled orders from non-employees, making the kit even more desirable for collectors. Current eBay listings show a Buy It Now price between $500 and $700, which is significantly higher than the $230 that the company asks for on the official website. However, X user @lithos_graphein shared a recreation of the kit available on Rebrickable, which added some tasteful modifications.

“A friend of mine fell in love with the ASML employee-exclusive Lego TWINSCAN EXE:5000 set and asked me to recreate it. While the original Lego version already looked impressive, I made some modifications to better capture the details of the real machine,” NightHawk11991 said in the description of the MOC (My Own Creation). “Using photos my friend took of the actual equipment, I designed this custom MOC to enhance its accuracy. I also added some interior details, including vacuum pumps, a laser emitter, reflectors, and more. Additionally, the silicon wafer tray can slide out for added functionality!”

The ASML TWINSCAN EXE:5000 MOC requires 152 unique parts and a total of 810 bricks. While this may not be the biggest build you can get from Lego or find on the Rebrickable MOC website, it’s still a rather significant build, with @lithos_graphein saying that it took them about a week to complete it. You can build and complete this project yourself, with PDF instructions available for free on the Rebrickable website. If you’re missing a Lego part to finish the build, you can also purchase it separately, although you might have a bit of a hard time looking for rarer parts included in the project.

@lithos_graphein is also the same one who was able to get their hands on a rare ASML Special Edition Monopoly board, which was released in 2007, after they traded an official ASML Lego kit for the board game. Company-commissioned merch like this is often given or offered to employees, meaning the average shopper wouldn’t be able to purchase it brand new unless they have a relative working in that company who buys it for them. Other semiconductor companies that have sold merch like this include TSMC and SK hynix, with resellers who were able to get their hands on company swag and sell it online for as much as $16,700.

Being unable to procure EUV litho tools from ASML, China must develop such systems locally, so from time to time we learn about an EUV-related 'breakthrough' in China that will turn into an actual EUV scanner over the following several years. This week, Chinese startup Langdao Silex demonstrated a prototype extreme ultraviolet (EUV) light source with over 15W of output power, 90% domestic components, and decent uptime. The company will need another three to five years to commercialize its technology, reports TrendForce, citing various sources. Yet, a light source alone does not make an EUV scanner.

An LPP EUV source from China

Langdao Silex, which was established in 2024 by Tsinghua University's Gas Discharge Plasma Laboratory alumni, demonstrated its EUV light source prototype at the 26^th China International Industry Fair in Shanghai, for obvious reasons, without lighting it up. Just like modern light sources from ASML, the company's low-power FlintNano LPP-OXXO system uses laser-produced plasma (LPP) technology that produces EUV radiation by bombarding tiny heated metal droplets with high-energy laser pulses to create extremely hot plasma. The system features a 1,000W driving laser (presumably a CO₂ one) with an operating frequency of 1 kHz – 20 kHz and conversion efficiency exceeding 2%. Its EUV output exceeds 15W (presumably measured at the plasma rather than at intermediate focus, as the company has not demonstrated a complete EUV collection system) while its brightness reaches 100 W/(mm²·sr). The source has an étendue of 1 × 10⁻³ mm²·sr and generates plasma with a diameter of 60 μm.

According to Langdao's specifications, the FlintNano LPP-OXXO can operate continuously for over 100 hours with plasma stability of 3% RMS and EUV output power stability of 5%. These characteristics make the system suitable for semiconductor exposure experiments, although its output remains insufficient for high-volume chip production. Meanwhile, the 'over 100 hours' availability is very impressive for a pre-production light source, though we must admit that there is no way to verify the claim.

Langdao is also developing the FlintMicro-LPP high-power EUV source, which has completed proof-of-principle validation, according to reports. However, Langdao has not disclosed its output power, conversion efficiency, or operating lifetime.

On paper, Langdao's EUV source is comparable in output to some of the earliest EUV sources developed by Western and Japanese equipment makers, albeit with a major caveat: Langdao seems to measure EUV power at the plasma rather than after the collector or at intermediate focus as ASML does. ASML's early EUV demonstration equipment used approximately 10W at intermediate focus DPP sources in 2007, while Cymer's LPP sources installed in early NXE:3100 systems delivered up to 11W during wafer exposure in 2012 (note that ASML's machines are source agnostic, so some NXE:3100 prototypes used DPP sources, while others used LPP sources). By 2013, Cymer demonstrated approximately 40W at intermediate focus over extended operating periods. Meanwhile, Japan's Gigaphoton targeted 25W with its first prototype in 2013 and demonstrated 43W in early 2014. By late 2014, Gigaphoton had achieved 118W during a 10-minute test. However, since these historical figures involve different measurement conditions and operating durations, direct comparisons are vastly inaccurate.

Not quite there

Langdao Silex has not demonstrated its EUV light source in action, so all we know about it comes from the company itself. As noted, 15W power looks comparable to early EUV sources developed by Philips Extreme UV (DPP) and Cymer (LPP). Meanwhile, early NXE:3400-series machines came with 125W – 150W at intermediate focus sources, which was far from enough for high-volume manufacturing (HVM).

Modern EUV lithography equipment comes with considerably more powerful sources. ASML introduced 250W-class EUV sources for commercial production in 2018 – 2019 and has since achieved substantially higher output levels. Increasing output requires more powerful CO₂ lasers, highly accurate tin-droplet generators, effective debris mitigation, durable collector mirrors, and considerably better long-term reliability.

Langdao claims that 90% of its prototype's components are domestically sourced and that most critical parts were developed in China. However, the company has not explained whether the localization figure reflects component count, value, or another metric. Furthermore, its representatives reportedly acknowledged that increasing EUV output and improving long-term operational stability are the targets to be achieved in a bid to commercialize the source.

Not an EUV scanner

Yet, even without complete EUV scanners, Langdao's source could support EUV photoresist development, mask inspection research, development of EUV collection and illumination optics, and eventually experimental exposure testing.

Developing a complete EUV lithography scanner will require further advances in EUV collection and illumination systems, high-precision multilayer mirrors, wafer and reticle stages, overlay accuracy, and system integration. Meanwhile, all of these devices require a separate supply chain that hardly exists in China today.

Therefore, even if Langdao successfully develops its EUV source and increases its output to levels suitable for high-volume manufacturing, this would not necessarily mean that China could manufacture its own EUV lithography machines any time soon. Moreover, while Langdao's progress is impressive, it took ASML/Cymer approximately six years to increase the output of its LPP EUV source from 11W to 125W – 150W at intermediate focus, so there is no way that Langdao's LPP source could be used for anything beyond development tools in the next three to five years.

Meanwhile, Langdao has not disclosed where its reported 15W of EUV output is measured. If this figure represents radiation generated at the plasma rather than power delivered at intermediate focus, the source's usable output would be considerably lower than 15W when measured according to industry-standard practices. That said, it does not look like China got significantly closer to building its own EUV lithography scanners with the current announcement.