OpenAI CEO Sam Altman said during a staff Q&A meeting that its latest model, GPT-5.6, is available in limited preview to only a small group of customers handpicked by the U.S. government. According to The Information, the federal government, specifically the Office of the National Cyber Director and the Office of Science and Technology Policy, asked the AI tech company to stagger the release of its latest model. While Altman did not mention how long the delay for the general release of GPT-5.6 will be, he said in a memo that he hoped it would happen in a couple of weeks. In the meantime, the U.S. government is granting access to the latest model on a case-by-case basis only.

Despite OpenAI’s agreement to the delay, sources say that Commerce Secretary Howard Lutnick called Altman to warn him against releasing GPT-5.6 to the public without prior approval from government agencies. “We’ve made clear to the U.S. government that this is not our preferred long-term model and will work with them and others in industry to achieve a more sustainable approach for future releases,” the OpenAI chief said in the Thursday memo.

This wasn’t the first time that an American AI lab has delayed the release of its frontier model due to security concerns. Back in early April, Anthropic released Claude Mythos Preview to select key institutions first, allowing them to prepare for the general release of the powerful AI model. It eventually built Fable 5, a watered-down version of Mythos with built-in safeguards to prevent misuse, and released it in June 2026. However, the U.S. government disagreed with the company’s belief that it was a safer model and put both Fable 5 and Mythos on an export control list just three days after it dropped. This meant that foreign nationals, even those who work for Anthropic, are banned from accessing the model. Since the company cannot enforce compliance, it just decided to pull the model completely from the market.

The increasing advancement of AI models has the White House scrambling to prevent them from falling into the wrong hands. This is especially true as it continues to compete with rival China for supremacy. Although the U.S. has taken steps like export controls to slow Beijing’s progress, many industry leaders believe that it’s only a matter of time before the East Asian country catches up. So, even though the Trump administration initially promised that it would reduce regulations to help AI advance much more quickly in the country, President Donald Trump has changed his tune and signed an executive order earlier this month that asks U.S. AI labs to give the government access to their latest models 30 days before it gets a general release.

However, this move has got some industry experts concerned. “…this escalation of government intervention is nothing to celebrate. It is horrible for the broader AI ecosystem,” Head of AI Policy and think tank Abundance Institute and former FTC Chief Technologist Neil Chilson said in their blog. “Continued arbitrary, unexplained deployment of export control authority will make companies slow-walk new models, depriving the public of powerful new tools. Every AI model, like all software before it, will have vulnerabilities that require patching. The US government should not hang a Sword of Damocles over every lab’s head, with no indication when it might drop or why.”

Meta has suspended an internal AI training program after an internal data leak exposed sensitive employee information company-wide, according to a Business Insider report on June 22. The program, introduced in April, was designed to help Meta train AI systems on real employee workflows by gathering data, but has now triggered internal backlash over privacy and data security.

Screenshots obtained by Business Insider showed that data collected through the program was more broadly accessible within Meta than intended. The exposed information reportedly included private employee conversations, performance-related data, transcriptions, and activity records. Meta classified the incident as a SEV 2, on an internal scale of 0 to 5, where SEV 0 is the most severe.

A Meta spokesperson confirmed that the company has paused the program while it investigates the incident. "We have carefully designed this program with privacy safeguards, and while we have no indication at this time that any data was improperly accessed by Meta employees, we're pausing it while we investigate," the spokesperson told Business Insider. The incident does not appear to be an external hack but rather an internal data mismanagement.

Meta introduced the program, called the Model Capability Initiative, to monitor employee behavior for use in improving its AI models. The program, which the company reportedly made mandatory for most staff, collected data on employees’ work activities, including keystrokes, mouse movements, conversations, transcripts, and performance-related information.

Employees were reportedly uncomfortable with the idea of their keystrokes and mouse movements being recorded for AI training. Now they're finding out the data may not have been properly protected, and was widely accessible across the company rather than restricted to intended viewers.

Screenshots reviewed by Business Insider reportedly showed employees criticizing the failure to lock down the data from the start. “I am incensed,” one employee wrote in an internal group, according to the report. Another said there was no evidence of malicious access, but called the lack of promised restrictions “super frustrating.”

The episode is the latest in a frustrating stretch for Meta's workforce. The company has cut thousands of jobs in part to fund AI infrastructure behind more powerful AI systems; the same class of systems that Meta and other companies are deploying to replace workers. Building these models also requires vast amounts of training data, and Meta turned to its own employees to supply it, a move most employees were reportedly against. Now these employees have learned that the data they were compelled to hand over was not adequately secured, leaving it exposed to much of the company.

The U.S. Department of Justice (DOJ) has filed a memorandum in response to NAACP’s lawsuit against Elon Musk’s xAI (now SpaceX) Colossus 2 data center. According to Wired, the federal government, through the DOJ, said that stopping the natural gas turbines needed to run the xAI data center “threatens American national, economic, and energy security by seeking to shut off the power supply for artificial-intelligence innovation that supports the Department of War’s military operations.” It further said that Grok is one of only four AI models the military and security agencies use to “support mission-critical operations across Secret and Top-Secret classified networks.”

Due to those purported national security interests, DOJ lawyers have joined xAI and the state of Missisippi in requesting that the lawsuit be dismissed.

The NAACP, through the Southern Environmental Law Center (SELC), sued xAI last year after an investigation discovered that Musk’s Colossus supercomputer facility used “illegal” generators to power its AI GPUs. The Memphis supercluster was launched in just 19 days — a major feat given that most projects of this scale usually take four years, according to Nvidia CEO Jensen Huang.

The site officially powered on in July 2024, but it wasn’t May 2025 that it became fully operational after it received 150MW of power from Memphis Light, Gas, and Water (MLGW) and the Tennessee Valley Authority (TVA). But instead of waiting months to get connected to the grid, xAI bridged the gap with mobile generators to get the electricity needed to run the hundreds of thousands of power-hungry GPUs.

It turns out that xAI failed to secure the permits needed to run the majority of these units. It’s been reported that the company applied for 15 portable turbines, but thermal images show at least 35 units on site. The company is allegedly using a loophole that allows it to run these units for 364 days without needing paperwork, but the Environmental Protect Agency (EPA) has since closed this loophole.

Aside from the offending generators in the first Colossus site, which is rented out to Anthropic, 27 natural gas turbines at Colossus 2 were cited as noncompliant in the initial lawsuit filed in April 2026.

However, as of May 2026, the SELC says that 57 turbines on the site are operating on site without a permit. According to the nonprofit, this resulted in a 111% increase in nitrogen oxide exhaust, an 83% jump in PM2.5 pollutants, and an 88% uptick in formaldehyde emissions since these generators were added. The lawsuit argues that these emissions endanger public health, and that continued use of the turbines “increases risks of asthma attacks and heart disease” in the surrounding communities.

In response to the DOJ's filing, the SELC issued the following statement: “With this filing, the Trump administration is launching an unprecedented attack on the public’s ability to defend themselves from illegal pollution. This is a blatant attempt to let well-connected corporations like xAI unlawfully pollute without any consequences, putting communities across the country at risk and threatening to open the door to large-scale pay-to-pollute corruption in the process,” SELC Litigation Director Kym Myer said. “The Department of Justice’s frivolous arguments fly in the face of decades of well-established legal precedent and we look forward to fighting them in court.”

The case is National Association for the Advancement of Colored People et al v. X.AI Corp. et al, filed in the United States District Court for the Northern District of Missourt.

Oracle's Project Jupiter data center is definitely an ambitious one, as its area spans 1,400 acres, or around 1,601 football fields, using a standard measuring unit. The build location is Doña Ana County, a rural desert area in the state of New Mexico with limited water supplies. As has become a common theme lately, the data center's water usage has become the focal point of a heated debate between residents, the county, and the developers, as detailed in an NPR report.

Residents are naturally concerned that an already-scarce resource will be drained further still, especially considering that the area's underground water level has been steadily dropping in recent years, forcing farmers to dig deeper wells to find a supply. Furthermore, the region is one of many affected by a Supreme Court decision meant to avoid draining the Rio Grande. New Mexico must reduce groundwater consumption by 5.9 billion gallons per year for 10 years. In Doña Ana and other affected areas, this means planning out the retirement of farmland, with all the economic and social impact that entails.

With all that going on, plus the overall zeitgeist around AI data centers, Jupiter's buildout generated quite the discussion. Most of the residents' concerns appear to come from the older build projects that included a power generation setup comprised of natural gas turbines and diesel generators, estimated by third parties to require at least a million gallons of water a day. However, Oracle claims it dropped that plan entirely in late April, turning instead to a solid-oxide fuel cell power plant that also has far lower carbon dioxide emissions.

Regarding the cooling for the data center itself, the firm has bought the water rights from a sod farmer who reportedly wasn't using them in full. Oracle says it's going to use a closed-loop system that will require a one-time fill of around 11 million gallons of water, spread across four buildings, and a maximum of 4,000 gallons per year for top-offs. The fuel cell system seemingly needs a 1-million-gallon initial fill and 168,000 gallons a year for top-offs, which Oracle says is less than the yearly water consumption of two average households. In an op-ed about the situation, the company calls these figures "negligible."

For once, the company may actually have a point. Those figures sound large in isolation, but according to a report for 2020, the county's farming used 410,000 acre-feet of water, equivalent to about 366 million gallons a day. The remaining daily usage for the county works out to approximately 45 million gallons a day. Although there's no shortage of stories about data centers draining local water supplies, their global usage is a figurative drop compared to activities like mere landscaping and golf courses, let alone industrial and farming use.

Oracle's revised plans and figures are all pretty recent, however, and Doña Ana residents claim they were barely given any time to properly analyze the proposals, or offered information about specific areas of impact for the drainage, particularly when the project proposal kept changing throughout the past year. State Rep. Micaela Lara Caden criticized the process, stating that it was "a $165 billion vote on a proposal that we knew about for less than a month".

In a council meeting last year, attendees asked the vote to be delayed, only to see four out of the five commissioners giving it a yay, prompting shouts of "recall!" by the audience. Susana Chaparro, the commissioner who voted "no," still holds reservations to this day, a sentiment shared by environmentally-minded people in the county. A local chile farm owner believes the county needs the project, but reportedly described the manner in which the county presented the project as "a fiasco."

Doña Ana is an embattled region, however, and Oracle pledged $360 million for schools and infrastructure, $50 million to upgrade a water utility that was found not to have been filtering out arsenic, and $12 million straight into the county's coffers. For a modest county with just over 220,000 residents and 40 settlements lacking paved roads and sewers, these figures represent a nigh-believable windfall. Manny Sanchez, the county commission's chairman, reportedly stated that "we've never had that type of money here in Doña Ana County."

A Meta data center project is reportedly causing issues in Morgan County, Georgia, where the community’s water supply has turned turbid right after it was constructed. Rep. Alexandria Ocasio-Cortez (D-N.Y.) showed the muddy water to the Environmental Protection Agency (EPA) Assistant. Administrator for Water Jessica Kramer during a congressional subcommittee hearing, asking if the office is looking into how data centers affect water quality across the nation. You can see the exchange in the embedded post below.

AOC: This is what drinking water in Georgia looks like after Meta began data center construction in the community.

— @acyn.bsky.social (@acyn.bsky.social.bsky.social) 2026-05-22T15:02:01.633Z

“This is the current drinking water in Morgan County, Georgia, right after a data center was constructed, the Meta data center was constructed. The only difference between the clean water and this was that data center,” Rep. Ocasio-Cortez said during the hearing while showing the jars of dirty water. She also added, “These families now have to ship, in a rural area, have to ship water to their house in order to cook and bathe themselves. Now I'm curious if the EPA plans any investigations on how data centers are affecting water quality and availability. I understood what you said about the rule, but are there going to be any open investigations on this issue?”

Kramer, for her part, said that she will investigate the issue as soon as she gets back to her office, saying, “It is a priority to ensure that water quality standards established by EPA are being met. And so, we'll be looking into that, certainly.”

A video by More Perfect Union, a progressive media and advocacy group, on YouTube Shorts showed that the residents in this rural part of Georgia use private and community wells, taking water directly from under the ground. However, their only source of water has allegedly become undrinkable since the data center construction began, and they also pointed to a significant groundwater recharge area, where water from the surface collects underground. The county government protects these areas as surrounding communities rely on them as a water source, and the nearby areas must be managed to minimize the threat of pollution.

It’s currently unclear what caused the water quality to drop in Morgan County. While it happened after the Meta data center construction began, correlation does not automatically equate to causation. One possibility is that the turbid water coming out of the wells of Morgan County residents is a sign that the water table is getting too low, and their pumps are now taking in water from the well bottom, where mud and sediment settle. When you combine that with reports of data centers using up a ton of water, then it is understandable that some in the community would blame the data center project for their water woes.

This also isn’t the first time that a data center project has allegedly created a water supply issue. Another massive project located in Georgia has reportedly secretly siphoned off 29 million gallons of water in a year and three months, resulting in low water pressure for residents. There have also been many concerns about these sites driving higher electricity prices for large swathes of the U.S., as well as possibly cutting out the lights for an entire town. We’ve also seen noise complaints coming from these developments.

All these issues have got many people complaining about these developments, with 7 out of 10 Americans now opposing data center developments near their homes. And even as AI tech companies and hyperscalers are pushing for more compute, more jurisdictions are passing data center bans all over the country.

Following a last-minute deal yesterday between Samsung Electronics and its South Korean workers’ union, the company will reportedly distribute 40 trillion won ($26.6 million) in bonuses to chip employees. According to a Bloomberg report, the bonus, which will vary by employee, will result in an average payout of 513 million won (~$339,000) per employee in Samsung’s semiconductor division, based on proposed terms and projected 2026 operating profits. Other estimates put the figure at around 600 million won (~$396,000).

As part of the tentative deal — which narrowly averted a strike originally scheduled to begin May 21st — Samsung has reportedly agreed to distribute 10.5% of its profits as employee bonuses in the form of stock, plus another 1.5% in cash. The union initially requested 15%. Addressing another major union demand, the new bonus program will continue for 10 years — rather than a one-off payment — provided specified profit targets are met.

The union plans to vote on the deal internally over the coming week. Should the deal pass, employees will likely receive the bonuses in early 2027. The agreement permits employees to sell one-third of the shares right away, and the remainder in installments over two years, according to Bloomberg.

The deal comes after months of escalating labor unrest centered on employee bonuses. Labor unions demanded a revision of Samsung’s bonus structure that would allow employees to earn more, as the company’s profits surge amid the AI-driven semiconductor boom. Bloomberg estimates Samsung’s 2026 operating profits will multiply sevenfold to 330 trillion won (~$218 billion).

The unprecedented profits are projected as the AI infrastructure boom has transformed memory chips — once a cyclical commodity business — into one of the most lucrative industries on earth. Demand for High Bandwidth Memory (HBM) and other AI-oriented components has triggered what analysts are calling a semiconductor supercycle. The workers at the center of this boom argue that since they are active facilitators of the hardware, they deserve a share of the profits. Samsung is notably not the first major Korean chipmaker to reach this conclusion. Last September, SK Hynix settled with its own union to allocate 10% of annual operating profit directly to employees as performance bonuses for the next decade, while removing caps on bonuses.

The payouts themselves are extraordinary by almost any conventional corporate standard — but then again, so are the profits currently flowing through the AI semiconductor ecosystem. Reports suggest the payouts are having a broader impact on the industry. Jobs at Samsung and SK Hynix were already coveted in South Korea. However, with potential bonuses that can exceed the lifetime earnings of workers in other sectors, competition for roles is sky-high. The bonuses are also reshaping decision-making within the companies themselves. We recently covered reports that Samsung and SK Hynix employees were considering terminating prestigious overseas training opportunities to remain eligible for bonuses.

Bandera, Texas, a small town of around 900 residents located about 40 miles northeast of San Antonio, Texas, opted to cancel its Flock AI contract. Three of the five-member town council voted to end the AI company’s services, after months of complaints from people who are wary about AI-powered government surveillance. However, one of the two people who wanted the security program to continue has publicly crashed out: According to 404 Media, councilor Jeff Flowers proposes that the town go back to a pre-digital age if its people want complete privacy.

“For months, I have listened to the outcry regarding License Plate Recognition (LPR) technology. I have seen the eyerolls, and I’ve even been met with ‘Nazi rhetoric,’ the dangerous claim that believing in accountability and community safety is somehow equivalent to totalitarianism. Comparing a neighbor’s desire for a safe street to a dark chapter of history is a classic case of comparing apples to oranges; it is a distraction used to avoid the reality of the threats our town faces today,” Flowers wrote in the local newspaper.

He also added, “Let’s take Bandera back to 1880 properly. No double standards, no hypocrisy. If LPRs are ‘unconstitutional’ and invade our right to ‘public’ privacy, we need to be courageous enough to go all the way. I look forward to the ‘Privacy First’ crowd showing up to support these bans [...] just remember to leave your phones at home.”

The town received a grant from the state of Texas to have eight Flock Safety AI license plate readers installed, but it turns out that its people do not want it. One resident said during a town hall that the community did not vote on the installation of these cameras, with another saying, “We don’t need to implement mass government surveillance in our town.” Aside from the protests during town meetings, the cameras themselves were repeatedly vandalized and had their poles cut down. The town had to pay for these repairs, meaning keeping the cameras operational is costing it extra.

Flock’s AI-powered license plate readers are in the center of several debates about security and privacy across the nation. While law enforcement agencies have praised the central database system used by the company, it’s also the same feature that has many people concerned. It’s feared that the authorities will use the data gathered by these cameras for targeting minorities and protesters, especially as reports have surfaced that immigration authorities have allegedly accessed the data gathered by the company without informing local police.

Bandera isn’t the only municipality to have turned off their Flock cameras. TechSpot says that at least 53 other jurisdictions located in 20 states have rejected the surveillance system in just the past six months. Flowers is framing the townspeople’s rejection of the AI-powered LPR system as a pushback against technology — but residents’ concerns aren’t just about privacy: they’re also about trust.

Samsung Electronics has narrowly avoided an 18-day strike at its South Korean chip operations after reaching a last-minute tentative wage agreement with its labor union on Wednesday — just hours before workers were due to walk out.

The union, representing nearly 48,000 members, said it would suspend the general strike planned for May 21 to June 7 and put the deal to an internal vote. Voting is expected to run from May 22 to May 27, though some notices put the window at May 23 to May 28. If members approve it, the tentative agreement becomes formal; if they reject it, the strike threat could return.

The deal followed last-minute talks involving Samsung management, union leaders, and South Korea’s labor ministry. Labor Minister Kim Young-hoon helped restart negotiations after earlier discussions broke down. Samsung’s Device Solutions negotiator, Yeo Myeong-gu, and union leader Choi Seung-ho signed the tentative agreement in Suwon.

Exact terms of the agreement have not been fully disclosed, but the dispute centered on performance bonuses. The union had pushed for Samsung to allocate 15% of annual operating profit to employee bonuses and scrap a 50% annual salary cap on performance payouts. Samsung had resisted those demands, arguing that profit-sharing across its businesses was complicated because its booming memory division and weaker logic/foundry operations were performing very differently.

The breakthrough caps months of escalating labor unrest. Dissatisfaction grew among workers as Samsung’s semiconductor profits surged on AI-driven memory demand, with workers comparing their payouts to those at rival SK Hynix, which has offered more generous bonus structures.

After earlier talks collapsed in February and March, the union threatened a May general strike. Further negotiations were scheduled for May 11 and 12, but the penultimate talks broke down again on May 12 without a deal. Leaked transcripts from the talks alleged that Samsung offered memory workers a 607% bonus worth $477,000, while logic chip staff were to get as little as 50%. The union rejected the proposal — citing a potential retention crisis due to the imbalance — and threatened to proceed with the strike.

The union had already demonstrated its ability to disrupt operations via a one-day strike that reportedly saw over 40,000 workers participate and caused night-shift output at Samsung’s foundry operations to fall sharply, with memory output also hit.

Ahead of today's deal, the May 21st strike seemed inevitable, with Samsung reportedly winding down chip production in preparation. Samsung had also obtained last-minute court injunctions to limit the impact of the strikes before direct government intervention helped bring both sides back to the negotiating table.

Analysts and officials warned of potential disruptions to chip supply and major economic damage in South Korea. One estimate cited possible daily losses of around 1 trillion won if production halted, while others warned of tens of billions of dollars in broader economic risk.

For now, that risk is suspended — not eliminated. The immediate question is whether Samsung’s union members accept the tentative wage deal. Approval will defuse the company’s biggest labor crisis in years, while a rejection will put the strike back on the table at one of the most sensitive points in the global AI chip supply chain.

Samsung and SK Hynix employees are reportedly rejecting overseas training programs in favor of bonuses while seeing their matchmaking grades rise on online dating sites. According to the Korean Media outlet Chosun, discussions on Samsung Electronics’ internal bulletin board indicate that employees are considering choosing potential bonuses over hard-earned, typically coveted training programs. The sentiments come amid projections that the company plans to distribute “record-high” performance bonuses due to profit surges from the AI-driven semiconductor boom, with the company’s policies restricting eligibility.

According to the report, Samsung Electronics has exempted employees currently undergoing overseas training from bonus eligibility. This stance is reportedly seeing a number of employees terminating their training to return to their roles. Samsung runs a highly sought-after training program that provides employees around their 7th year at the company with financial and operational support to pursue postgraduate degrees at prestigious institutions both in and outside Korea. For example, in addition to their annual salaries, the company provides employees pursuing an MBA in the US with 500 million Korean won (~$333,000) in tuition and living expenses over two years. The support varies across institutions and countries.

The program is typically highly competitive, with only 1 in 70 applicants succeeding. Now, following projections that the electronics giant could distribute an average of 600 million (~$400,000)won in bonuses to employees in the semiconductor division, interest in the training programs has dropped. Reports suggest that even those currently undergoing training are considering halting midway to be eligible for performance bonuses, as these are likely to outweigh the financial benefits of the training program.

According to the Chosun report, sources from Samsung Electronics stated, “There have been inquiries to the HR team about whether trainees can withdraw mid-program and if they must repay previously received company support.” indicating that despite potential repayment obligations, some employees are willing to return regardless.

These reports are coming even as Samsung Electronics employees are planning industrial action to compel the company to extend continuous performance bonuses to employees outside the semiconductor division. The company says only employees in the semiconductor division will receive ongoing bonuses, while others in other divisions will receive a one-time payment. The strike is scheduled for the 21st of May.

At Samsung’s competitor, SK Hynix, similar reports of employees abandoning training in favor of bonuses are emerging. Despite the fact that SK Hynix takes a softer stance by offering partial bonuses to training beneficiaries, anonymous posts on its internal board reportedly include comments like, “If next year’s bonus reaches 700 million Korean won (~$466,000), not abandoning the training might make me a sudden pauper overnight” and “Seeing the projected bonuses for next and the following year, I want to cut off the hand that applied for the training.”

These comments follow projections of even bigger bonuses in the coming years, as SK Hynix paid out 140 million Korean won (~$93,000) in performance bonuses early this year, which, unlike at Samsung, was distributed across the entire company. Analysts predict a dip in the popularity of overseas training programs at Samsung Electronics and SK Hynix over the next few years, as the appeal of performance bonuses grows. SK hynix bonuses are tipped to hit $477,000 this year and as much as $900,000 next year.

Beyond strikes, life-changing decisions, and financial gain, these massive performance bonuses stemming from AI-driven profit surges are influencing quite unexpected areas of life in South Korea. Reports from Chosun indicate that SK Hynix employees are ranking higher on matchmaking companies’ eligibility scales due to bonuses and increased income, with employees reportedly being flooded with blind-date offers.

"In the past, if we matched a partner for an SK hynix employee at about a B+ grade, now it's unconditionally 'A-grade.' As the overwhelming bonus system has become known, we're seeing a trend where female members are seeking out SK hynix employees first,” said Son Dong-gyu, CEO of the matchmaking company Bien Aller.

However, on the flip side, because bonuses are paid in direct proportion to days worked, there has been an uptick in aversion to taking extended leave, such as parental leave, as these are reportedly perceived as forfeiting income. These reports all underscore the massive influence of the AI boom on almost every aspect of humanity.

Arm Holdings, the maker of the popular Arm architecture used by Qualcomm, Apple, and several other companies, is facing an antitrust investigation as the U.S. Federal Trade Commission (FTC) looks into the company’s operation. Bloomberg reports that the FTC is determining if the company is trying to monopolize the Arm architecture and either only give customers lower quality designs for their own semiconductors or outright deny them access to its licenses. The move comes as Arm launched its own AGI CPU focused on data centers in March 2026, a significant departure from the company, whose business previously focused on licensing its chip designs to other companies.

The company’s legal troubles began when it sued Qualcomm, the biggest manufacturer of smartphone chips, for using Nuvia’s ARM licenses after it acquired the startup in 2022. Arm Holdings argued that the smartphone chipmaker cannot use Nuvia’s licenses after its acquisition, and that it should have acquired a new one to continue using the startup’s designs based on Arm licenses. Arm ultimately lost the case, allowing Qualcomm to continue using the Oryon cores it acquired from Nuvia.

The Qualcomm lawsuit broke the longstanding relationship the two companies had, and it also opened a can of worms for Arm. The former launched a global antitrust campaign against the latter because of the case, saying that it was using its dominant market position to prevent competition. It reached out to the European Commission, the U.S. FTC, and Korea’s Fair Trade Commission to present its case, with the Korean government agency raiding Arm’s Seoul office in November 2025.

While the x86 processor still has the advantage when it comes to desktops and laptops, its advantage is slowly being eroded by Arm-based Apple Silicon and Qualcomm’s Snapdragon X-series chips. On the other hand, it has cornered the mobile market, with Apple, Qualcomm, Samsung, MediaTek, and other mobile device chipmakers using Arm architecture. Some analysts also say that it will eventually dominate the AI server industry, with over 90% of custom processors expected to use an Arm chip by 2029.

Arm’s launch of its own AGI processor has got its customers (and now potential competitors as well) wary of its status as both chip designer and manufacturer. Arm did not introduce a processor for consumer computing, meaning it hasn’t directly challenged some of its biggest customers. However, the fact that it has expanded into building physical chips has worried them that it could leverage the popularity of the Arm architecture to gain an unfair advantage and limit competition.

Intel has announced a multi-year strategic partnership with F1's legendary McLaren Racing team. The chipmaker is now the Official Compute Partner of the McLaren Mastercard Formula 1 Team, Arrow McLaren IndyCar Team, and McLaren F1 Sim Racing Team, according to a press release today. Interestingly, this announcement again pits Intel against its PC chip industry nemesis, AMD. The Red Team has already been working in partnership with the Mercedes-AMG Petronas Formula One Team for six years.

Intel's PR blurb says F1 racing is "one of the world’s most technologically demanding sports." Thus, Intel engineers will be tasked with delivering advanced computing for AI and high-performance architectures that are required to keep McLaren competitive.

The new agreement means that systems using Intel Xeon and Core Ultra chips will be leveraged to support McLaren's quest for the ultimate performance on the track. Specific calculations that F1 engineers spend their days optimizing for include "performance-critical workloads, including computational fluid dynamics, aerodynamic analysis, vehicle-dynamics simulation, [and] race strategy analytics." As well as real-time data, F1 support computers are used to sift through massive volumes of post-race data. Computers used to optimize F1 racing cars are also increasingly using AI tools, plus low-latency edge computing solutions, and diverse software platforms.

"Formula 1 racing and IndyCar are some of the ultimate proving grounds for high-performance computing. Intel is proud to be McLaren Racing’s compute partner, and to be part of a team that thrives on precision, speed, and innovation," said Lip-Bu Tan, Intel CEO. "Together, Intel and McLaren will push the boundaries of what’s possible, transforming data into competitive advantage at every turn."

A statement by Zak Brown, CEO, McLaren Racing, confirmed Intel hardware had already been an important part of the F1 team's tech ecosystem. It will be interesting to see if the new, closer relationship will produce noticeable results on the circuits around the globe, burning rubber at speeds pushing beyond 230 mph.

As we mentioned in the intro, AMD has already been working closely with a major F1 team for years. The firm has a page dedicated to how AMD Epyc and Threadripper processors are a competitive edge for the Mercedes-AMG PETRONAS Formula One team. Similar to Intel's announcement today, the AMD/Mercedes partnership uses advanced compute for "aerodynamic simulation and faster data analysis."

AMD's partnership with Mercedes-AMG was forged back in 2020, which might explain why there's no mention of artificial intelligence in the linked PR blurb, yet.

Amazon is the latest hyperscaler where employees have been caught inflating AI token consumption to hit internal usage targets, following similar behavior documented at Meta and Microsoft last month, the Financial Times reports.

The company set targets requiring more than 80% of its developers to use AI tools each week and tracked consumption on internal leaderboards. Some employees told FT they had been using MeshClaw, an in-house agent platform that can initiate code deployments, triage emails, and interact with Slack to maximize their token numbers. Amazon said usage statistics would not factor into performance evaluations, but multiple employees said they believed managers were monitoring the data. One said there was "so much pressure to use these tools," another described how tracking created "perverse incentives."

The practice — dubbed "tokenmaxxing" — has become widespread enough to generate its own vocabulary and leaderboards, but beyond workplace culture, if a meaningful share of AI consumption is performative, how reliable are the demand figures that hundreds of billions in AI infrastructure procurement are being allocated against?

Combined 2026 capex from Amazon, Microsoft, Alphabet, and Meta is tracking between $650 billion and $700 billion, with some Wall Street projections exceeding $1 trillion for 2027, and every hyperscaler has told investors that inference capacity is being absorbed as fast as it can be deployed. Internal developer consumption is obviously part of that absorption, and it sits alongside paying external customers in the usage data that informs the likes of capacity planning, GPU orders, HBM procurement, and power infrastructure.

Tokenmaxxing doesn’t mean the demand is fabricated — enterprise AI adoption is broadening, and inference workloads are scaling into production — but there’s a distinction between adoption and consumption intensity. The former is a durable driver of demand, whereas the latter is gameable, and it’s currently being amplified by the incentive structures that these companies built. The water is further muddied by reports that AI is more expensive than actual workers.

Meta's internal leaderboard lasted days after public exposure, and Amazon recently restricted visibility of team-wide usage statistics. And when measurement shifts, the consumption intensity they incentivized will shift with them.

Nvidia CEO Jensen Huang has highlighted per-engineer token consumption as a key metric, stating he’d be "deeply alarmed" if a $500,000-a-year engineer was not consuming at least $250,000 in tokens. Nvidia's inference growth obviously depends on that consumption being a productive workload that persists and compounds because every inflated token is real GPU time.

Angie Jones, formerly VP of engineering for AI tools at Block, told LeadDev she expected the industry to pivot toward measuring efficient token usage rather than celebrating volume. In a cycle where GPU orders and power commitments are being placed years in advance, the quality of the demand projections behind them matters. The hyperscalers are building for a world where every knowledge worker consumes hundreds of thousands of dollars in annual compute. Whether that consumption proves productive or performative will determine how much of this year's $700 billion generates durable returns.

Jobs cuts related to increased AI usage have practically become background noise at this point, but there's still the occasional outlier that merits inspection. In this case, Cloudflare announced both its quarterly results and the fact that it's cutting 20% of its workforce over this year — to the tune of 1,100 heads.

Although Wall Street investors are usually amenable to job cuts, as they're generally seen as source of added short-term profit, this time around the market didn't react kindly at all, tossing Cloudflare's stock price down a cliff to the tune of a 19% drop in one day alone — nearly a 24% drop for the week. In a tweet, CEO Matthew Prince noted that "very few engineers or customer-facing sales people" are impacted by the layoffs, and that the firm wants to "continue to hire like crazy" for those roles.

The firm expects to take a $140-$150 million charge associated with the job cuts, and the soon-departed also ought to receive equity in the company as of their professional passing. The Q1 2026 results actually matched or beat both its revenue and Earnings Per Share (EPS) estimates — with a revenue of $639.8 million (up 34% year-over-year) and an adjusted EPS of $0.25 (beating the expectation of $0.23). Also, it's worth noting that Cloudflare is still up around 30% since the start of the year.

It's hard to pin down why investors are so disappointed, but perhaps the reason is that Cloudflare issued Q2 2026 guidance stating that its revenue will be around $644.5 million — a figure described as "just shy" or "below" estimates, thus displaying a lack of growth. Stockholders may have gotten the impression that the company needed the job cuts for financial growth in this environment, but, as the old adage says, markets will do market things.

As befits any harsh breakup, Cloudflare's management sent out a long message to employees detailing the why and how of the move. The company claims it increased the usage of AI "by more than 600% in the last three months alone," noting that most every part of the company runs "thousands of AI agent sessions" daily. Thus, Cloudflare believes it "[has] to be intentional in how we architect our company for the agentic AI era," and notes that the decision is "is not a reflection of the individual work."

U.S. District Judge Jon Tigar has denied Nvidia's request to dismiss a copyright infringement case filed against it, arguing that it’s not liable for how clients use its AI-powered NeMo Megatron Framework. According to TorrentFreak, Nvidia is asking the court to dismiss the direct copyright infringement claims that are connected to its use of the Bibliotik eBook torrent tracker, the Books3 dataset, and 'The Pile' dataset for language modeling. Nvidia then cited the Cox vs. Sony ruling, where the U.S. Supreme Court ruled that a service provider is not liable for any piracy that its users might carry out.

Nvidia said that its NeMo Megatron Framework has significant “non-infringing uses” and that it did not promote it as a piracy tool. This should fall under Justice Clarence Thomas’ decision saying, “Under our precedents, a company is not liable as a copyright infringer for merely providing a service to the general public with knowledge that it will used by some to infringe copyrights.” Unfortunately for the company, Judge Tigar disagreed with its argument, saying that it’s not the framework, but specific scripts within it that violated copyright rules.

He said that these were intended to make it easier for users to automatically download and preprocess The Pile dataset, which the complainants said allegedly contained copyrighted work. “The scripts are alleged to have no other purpose than to speed up the process of infringement, unlike the digital video recorder systems at issue in Sony Corp. or the internet service provided in Cox,” Judge Tigar wrote. Bibliotik is a private eBook torrent tracker, which allegedly contains over 197,000 books. It was then included in the Books3 dataset, which itself was included in the 800+ gigabyte The Pile dataset. The Pile was then used for training Nvidia’s AI LLMs, resulting in several authors filing a class action lawsuit against the company for copyright infringement.

There have been previous cases of copyright infringement related to AI companies scraping data for training their models. Aside from this case against Nvidia, Meta has also been facing a similar lawsuit since last year. It even defended itself by saying that using pirated material is legal if you don’t seed content. Google has even been pushing to have AI scraping tagged as fair use, saying that it wants “copyright systems that enable appropriate and fair use of copyrighted content to enable the training of AI models in Australia on a broad and diverse range of data while supporting workable opt-outs for entities that prefer their data not to be trained in using AI systems.”

With this decision, the authors’ class action against Nvidia is set to move forward, and we will likely hear more details as the case progresses. We don’t have a date yet for when the next hearing will be, though. Still, we expect this to be a multi-year battle as the AI giant battles it out with allegedly infringed writers.

Negotiations between the National Samsung Electronics Union, which represents workers in the company’s chipmaking division, and management have seemingly broken down over a single issue. According to the Financial Times, the two sides are close to agreeing on an allocation of 13% of operating profit, which works out to be roughly $340,000 USD per employee, as a bonus to the workers. However, company management is only willing to give this as a one-time offer, while the union wants the allocation to be guaranteed annually and included in the agreement that the two sides will sign.

The issue about the bonus was brought to national attention when more than 30,000 Samsung workers took to the streets in late April to demand a bigger slice of the profits that Samsung is making from the AI infrastructure buildout. This demand stemmed from a comparison with SK hynix workers, who were guaranteed bonuses of $477,000 each this year, and almost doubled to $900,000 next year. Furthermore, these bonuses are guaranteed for the next ten years. The massive amount stemmed from the windfall that the memory and storage chip manufacturer is making from the AI infrastructure build-out, with AI data centers and hyperscalers willing to pay a premium just to secure the chips they need.

The workers argue that even though Samsung is much larger, their bonuses only equate to less than 30% than what SK hynix offers to its people. Their initial demand was a 15% cut in the semiconductor fab’s operating profit, a removal of the 50% bonus cap, and a 7% wage hike, while management countered with a 10% allocation, a 6.2% pay increase, and other benefits like preferential mortgage rates. It seems that the two sides have finally settled on the 13% bonus allocation, and the only question remaining is whether management will agree to a guaranteed annual bonus.

Samsung is facing the threat of a massive strike if the two parties fail to come to an agreement. The union said that it will conduct a general strike from May 21 to June 7 — meaning Samsung’s chip fab operations will be crippled for at least 18 days. It should be noted that the single-day action in April resulted in a 58% drop in production for just a single shift. Prof. Kwon Seok-joon of Sungkyunkwan University told the Financial Times that an action like this could cost Samsung somewhere between $6.9 billion and $11.7 billion in direct losses, with an even larger amount in indirect costs. Furthermore, it will damage Samsung’s reputation as a supplier for HBM4 chips, especially with the tight competition between the three major memory and storage chip manufacturers.

Kwon also noted that it’s harder for Samsung to just grant the bonus, even if it comes from operating profits and not revenue. Unlike SK hynix, which is a standalone fab, Samsung’s semiconductor fab operates under the larger Samsung Electronics company, which itself is part of the larger Samsung Group. Other arms of the company are said to be struggling because of the higher costs brought about by the chip shortage, even as the semiconductor division is making record profits.

If the fab workers get their desired bonus, workers from less profitable divisions might feel that they’re being short-changed. Because of this, a smaller union, whose members mostly come from Samsung’s smartphone, TV, and home appliances lines, has reportedly pulled out of the planned joint strike. Should management grant the semiconductor division’s demand for a bonus allocation of 15% of operating profits, Kwon told FT that “the maths gets uncomfortable fast.”

Intel has announced a pair of high-profile leadership appointments as it doubles down on its AI ambitions. According to an official press release today, the company is bringing in industry veteran Alex Katouzian to lead its client computing and emerging “physical AI” efforts, while Pushkar Ranade is coming in as Chief Technology Officer to steer the company’s long-term innovation strategy. The appointments clearly signal Intel's intent to more tightly align its core computing business with the rapidly evolving AI landscape, particularly at the edge and in real-world systems.

Alex Katouzian is joining as executive vice president and general manager of the Client Computing and Physical AI Group, a role Intel says will align its traditional PC-focused business with emerging AI-driven systems spanning robotics, autonomous machines, and edge devices.

Katouzian is bringing 25 years of industry experience from Qualcomm, spanning from when he joined as a senior engineer in 2002 to his last very senior role as executive vice president and group general manager of mobile, compute, and extended reality (XR).

He is widely regarded as a key figure in scaling Qualcomm’s mobile and compute platforms globally and is responsible for significant contributions to the industry, including the expansion of Snapdragon mobile platforms and Qualcomm’s push into PC and XR computing. Intel believes this experience will prove invaluable as the company moves to expand beyond traditional PCs into AI-enabled, connected, and edge-based computing ecosystems.

Speaking on the appointment, which officially kicks off this month, Katouzian expressed excitement about joining Intel at what he described as a pivotal moment for AI-driven transformation across computing platforms.

“Intel is creating the foundation for AI-driven transformation, from leading in AI PCs, to scaling AI inference at the edge, and accelerating the future of physical AI systems,” wrote Katouzian. “I’m excited to join Lip-Bu and the Intel team at this critical moment to help scale innovation and deliver the next generation of computing experiences.” Katouzian announced his departure from Qualcomm on LinkedIn.

The second appointment sees Pushkar Ranade transition to permanent Chief Technology Officer after serving in the role on an interim basis for several months. Ranade is not a new face at Intel and has been with the company for over 10 years.

According to Intel, his role will involve leading the company’s technology strategy, overseeing special technology initiatives, and driving development across emerging areas such as quantum computing, neuromorphic computing, photonics, and advanced materials. He will also serve as chief of staff to the company’s CEO.

Both appointments continue an existing trend as more tech companies pivot to a stronger AI focus. Microsoft, for instance, created a dedicated AI division led by DeepMind co-founder Mustafa Suleyman, alongside a series of follow-on hires to strengthen its AI leadership. Google has also brought in senior robotics leadership to push AI into physical systems.

You’ve probably seen a few examples of PC DIYers installing figurines in their elaborate, RGB-infused fishtank builds. Chinese TechTuber Soda Baka has scaled this phenomenon up to create a tower PC large enough for a human to work and play inside. What looks like a figurine in some of these PC pictures is actually the TechTuber posing. The video is shared on China’s Bilibili.

We aren’t 100% sure of the purpose of this build, other than spectacle, due to things getting lost in translation. However, a short skit in the video (machine translated) indicates the build was influenced by Baka’s little cousin being barked at by his mom “You play computer games all day, why don’t you just live in a computer!” Something like that.

We also get the feeling this is a canny video sponsorship deal as air conditioners start to become highly desirable, now that we are moving into warmer months in the Northern Hemisphere.

The human-habitable PC tower project gets underway with Soda Baka sketching and modeling before the construction begins in earnest. We see various huge components being fabricated, that look like case fans, GPUs, AiO CPU coolers, RAM sticks and so on. The TechTuber doesn’t forget to add splashes of RGB lighting, of course.

Some real-PC stuff is installed in the mega-PC, so there’s something to do once the side panel is put on… You can see Soda Baka sitting at a compact desk playing a game on the screen attached to the AiO – how modern.

Now things heat up, literally, as to simulate PC-scale heat at this large scale some hot-coal sauna installations are used. We see the TechTuber boosting the heat/humidity levels with a bottle of water. Apparently it doesn’t take long for the enclosed computer room (in the PC) temperatures to rise above 100°F (38 degrees Celsius).

Because all the huge fans and coolers are fake, our trapped TechTuber has to turn on the installed 12kW AC unit, which boasts 820m³/hr air circulation to cool the ‘PC system.’ Quickly enough the atmosphere changes to become “like a spring evening in the North [of China].” Was it all an elaborate Midea 3^rd Gen Pro AC unit sponsorship promotional video? Maybe it was.

Google, Amazon, Microsoft, and Meta plan to spend a combined $725 billion on capital expenditure in 2026, a 77% increase over last year's record $410 billion, according to first-quarter earnings reports compiled by the Financial Times.

Google led with 63% cloud revenue growth and an 81% jump in net income to $62.6 billion, while Meta's stock dropped 6% after hours despite a 33% revenue increase, punished by investors for adding $10 billion to its spending forecast and offering no firm timeline on new AI models.

But in the earnings calls, at least two of the four companies explicitly blamed rising memory chip prices for pushing budgets higher, confirming what market data and industry executives have been warning about for months.

Memory costs inside the capex

Microsoft’s CFO, Amy Hood, told investors that rising prices for memory chips and other components accounted for $25 billion of the company's record capex budget. Microsoft set its 2026 spending at $190 billion, far above the $152 billion average analyst forecast. Hood warned that even with the additional investment, Microsoft expects to remain capacity-constrained on GPUs, CPUs, and storage through at least 2026.

Meta cited the same, with the company raising its full-year capex range to $125 billion to $145 billion, up from a prior ceiling of $135 billion. In its earnings release, Meta attributed the increase to "higher component pricing this year, particularly memory," alongside rising costs for land, power, and skilled workers needed to build data centers that now consume 70% of the world's memory output.

The timing of all this is hardly coincidental, with TrendForce having reported DRAM contract prices rising roughly 95% quarter over quarter in Q1 2026, with a further 58% to 63% increase projected for Q2. NAND is following a similar trajectory, with Q2 contract prices expected to climb 70% to 75%. Server DRAM and high-density DDR5 RDIMMs are absorbing the bulk of production capacity, and all NAND output for 2026 is already committed, according to Phison CEO Khein-Seng Pua.

Hood's $25 billion, therefore, helps to put a dollar value on what has previously been an abstract concern: If one company's memory cost inflation alone exceeds the entire annual capex of most semiconductor firms, the pressure on consumer DRAM and NAND supply becomes much easier to quantify.

Google Cloud's contract backlog

Meta and Microsoft aside, Google’s Cloud revenue hit $20 billion in the same quarter, growing 63% year over year, outpacing both Amazon Web Services ($37.6 billion, up $8.3 billion) and Microsoft's Azure-driven cloud segment ($34.7 billion, up $7.9 billion).

Google's cloud contract backlog reached $460 billion, roughly double the $240 billion reported at the end of Q4 2025. Amazon reported $364 billion in its own pipeline, which will expand further after a recent $100 billion computing contract with Anthropic over the next decade. Microsoft's commercial remaining performance obligations hit $625 billion, up 110% year over year.

Cloud boss Thomas Kurian attributed Google's growth to its strategy of building custom AI chips, foundation models, and products in-house, telling the Financial Times that this gives the company a cost and research advantage over competitors that have struggled to develop their own chips and frontier models. Google's 7th-gen Ironwood TPU, which packs 192 GB of HBM3E per chip with 7.37 TB/s bandwidth in pods of up to 9,216 chips, is central to that strategy, and Anthropic has committed to access up to one million of them. Google recently unveiled its 8th-gen TPUs, which are split into two distinct variants for training and inference.

Alphabet raised its capex guidance to between $180 billion and $190 billion, up $5 billion from its previous guidance of $175 billion. CFO Anat Ashkenazi said he expects capex to “significantly increase” in 2027, causing shares to rise by some 7% after hours. It’s worth noting that $37.7 billion of Alphabet’s net income of $62.6 billion came from unrealized gains on non-marketable equity securities, primarily the company's Anthropic stake, according to the earnings release filed with the SEC. Strip that out, and operating performance was still strong, with a 36.1% operating margin, but the total net income number overstates recurring profitability.

Custom silicon and the GPU question

These capex figures reflect more than GPU purchases, because each hyperscaler is now deploying or developing custom accelerators to reduce dependence on Nvidia for inference-based workloads.

Amazon's Trainium3, built on a 3nm process with 144 GB of HBM3E and roughly 4.9 TB/s of bandwidth, is what CEO Andy Jassy described as "nearly fully subscribed" for 2026, and Meta has announced four generations of its MTIA inference chip, all fabbed at TSMC alongside Broadcom, even as it signed GPU deals worth roughly $110 billion combined with AMD and Nvidia. Meanwhile,. Microsoft's Maia 200 is deploying in U.S. Central data centers.

This pattern is likely to extend beyond accelerators as CPU demand for agentic AI workloads drives a parallel supply crunch with CPU lead times currently stretching to six months. Intel has reported billions in unmet Xeon demand, and Arm CEO Rene Haas has stated that agentic workloads require roughly 120 million CPU cores per gigawatt of data center capacity, four times what traditional AI training clusters need. Per Intel CFO David Zinsner, data center CPU-to-GPU ratios have already moved from 1:8 to 1:4, with further convergence expected to reach or go beyond parity.

Despite record spending, all four companies have acknowledged supply constraints that additional capital alone can’t resolve. Nvidia has booked an estimated 800,000 to 850,000 wafers of TSMC's CoWoS advanced packaging capacity for 2026, consuming over half of the total output and leaving AMD, Broadcom, and Google's TPU program competing for the remainder. CoWoS remains oversubscribed through at least mid-2026, and TSMC's U.S. packaging fabs aren’t expected to reach volume until 2028.

Power infrastructure is another bottleneck, with large power transformer lead times extending to roughly 128 weeks, and the IEA estimating that approximately 20% of planned global data center projects could be at risk of grid-related delays. TrendForce recently downgraded its full-year server shipment growth forecast from 20% to 13% because power management ICs and baseboard management controllers needed to assemble complete servers are stretching to 35- to 40-week lead times. Samsung's planned closure of its S7 eight-inch wafer fab in Korea will tighten PMIC supply further.

‘The bear thesis is garbage’

Meta's stock slipped by 6% after-hours following the earnings, erasing roughly $113 billion in market value. That drop reflected both the $10 billion capex increase and CEO Mark Zuckerberg's lack of a firm schedule for releasing improved AI models to follow the recently launched Muse Spark. Dec Mullarkey, managing director of SLC Management, told the FT that investors are concerned about whether Meta's historically capital-light business is becoming far more capital-intensive.

"The bear thesis is garbage," countered Brent Thill, an analyst at Jefferies, arguing that revenue growth across the sector justifies the spending. Zuckerberg offered little to settle the debate. Asked about Meta's AI agent development, he told investors he cared more about quality than deadlines, adding that most AI agents available today are not good enough for everyday users.

Amazon kept its $200 billion capex plan unchanged, and Microsoft CEO Satya Nadella said ending his company's exclusive contract with OpenAI was beneficial, claiming royalty-free access to OpenAI's frontier models and IP through 2032.

Meta CEO Mark Zuckerberg told employees at a company town hall on Thursday that the roughly 8,000 planned layoffs are a direct consequence of the company's ballooning AI infrastructure budget, Forbes reports. The cuts, which affect about 10% of Meta's workforce and are set to begin on May 20, come the same week the company raised its full-year 2026 capital expenditure forecast to between $125 billion and $145 billion, up from a prior range of $115 billion to $135 billion.

“We basically have two major cost centers in the company: compute infrastructure and people-oriented things," Zuckerberg ​said during the town hall, as heard by Reuters. With more capital flowing toward AI hardware, he said, there is less available for headcount. He also declined to rule out further reductions later in the year. Meta spent $72.2 billion on capex in all of 2025. The midpoint of its new 2026 guidance would nearly double that figure in a single year.

The timing undercuts any suggestion that Meta is cutting jobs out of financial necessity. The company's Q1 2026 earnings, reported on Wednesday, showed revenue of $56.31 billion, a 33% increase year over year, while net income hit $26.8 billion. Q1 capital expenditure alone reached $19.84 billion, and CFO Susan Li told investors she couldn’t predict the company's optimal long-term workforce size given how quickly AI capabilities are evolving.

Zuckerberg's comments land in the middle of a growing debate about whether companies are using AI as a convenient justification for workforce reductions they would make regardless. OpenAI CEO Sam Altman raised the issue in February, telling CNBC at the India AI Impact Summit that some firms engage in "AI washing" by attributing layoffs to the technology when the actual reasons lie elsewhere.

Zuckerberg's explanation is more specific than most, explicitly pointing to infrastructure spending rather than AI-driven productivity gains as the driver, but that specificity raises its own tension. Nvidia’s VP of applied deep learning, Bryan Catanzaro, said earlier this week that compute already costs more than the employees on his team, and a 2024 MIT study also found AI automation was economically viable in only 23% of vision-related roles. If AI infrastructure is currently more expensive than the labor it supplements, the return on trading one for the other remains a dubious strategy.

Some employees have understandably criticized Zuckerberg and other executives on Meta's internal message board over the layoffs and a separate initiative to monitor employee productivity through mouse and keyboard activity tracking. "Getting everyone internally to use AI tools and getting to do the work more efficiently is not the thing that's driving layoffs," but added that Meta will monitor “how all this stuff trends.”

These are the largest cuts at Meta since the 2022 and 2023 rounds that shed roughly 21,000 positions. Across the broader tech industry, more than 80,000 workers have been laid off in 2026 so far.

Google, Amazon, Microsoft, and Meta collectively plan to spend $725 billion on capex in 2026, up 77% from last year's record $410 billion, according to first-quarter earnings compiled by the Financial Times. Google delivered the strongest results, with cloud revenue jumping 63% year over year to $20 billion, while rising memory chip prices pushed spending forecasts higher at both Microsoft and Meta.

"The AI economy is healthy," Brent Thill, an analyst at Jefferies, told the Financial Times, adding that recent revenue growth justified the enormous capital outlays. "The bear thesis is garbage."

Microsoft set its calendar-year 2026 capex at $190 billion, well above the $152 billion average analyst estimate. The company’s CFO, Amy Hood, attributed $25 billion of that figure to rising memory chip and component costs. She told investors that despite the additional spending, Microsoft expects to remain capacity-constrained through at least 2026 as it works to bring GPU, CPU, and storage infrastructure online faster.

Meta increased its full-year projection by $10 billion to a range topping $145 billion. The company cited higher component pricing, particularly for memory, alongside growing competition for land, power, and skilled workers needed to build data centers. Revenue grew 33% to $56.3 billion.

Dec Mullarkey, managing director of SLC Management, told the Financial Times that investors are growing uneasy with Meta's escalating infrastructure costs, questioning whether a historically lean business is becoming far more capital-hungry. “Investors continue to be concerned about how Zuckerberg’s once capital-light money machine may be morphing into a capital-intensive incinerator,” he said.

Alphabet posted an 81% increase in net income to $62.6 billion on revenue of $110 billion. Google Cloud reached $20 billion in quarterly revenue, growing faster than Amazon Web Services ($37.6 billion total, adding $8.3 billion year over year) and Microsoft's Azure-driven cloud segment ($34.7 billion total, adding $7.9 billion).

The company's cloud contract backlog reached $460 billion, roughly double the $240 billion reported at the end of Q4 2025. Google Cloud boss Thomas Kurian credited the company's strategy of building custom AI chips, foundation models, and products in-house for giving it a cost and research advantage over competitors. Alphabet's capex guidance rose by $5 billion to as much as $190 billion, matching Microsoft. Shares climbed 7% after hours, putting Alphabet on track for a record $4.3 trillion market valuation.

CEO Mark Zuckerberg offered no firm schedule for releasing improved AI models to follow the recently launched Muse Spark. Asked about the pace of Meta's AI agent development, Zuckerberg told investors: "There's a lot of agents out there that people are building for different things, but there aren't that many that I would want to give to my mother."

Researchers at Princeton University have created a three-dimensional neural network device that combines living brain cells and advanced embedded electronics. According to a recent press release, this 3D bioelectronic computer was programmed to differentiate patterns using computational techniques.

Basically, we are looking at living brain cells performing computational tasks outside the brain, using embedded electronics. This is not the first time scientists have used brain cells to perform computation. In previous attempts, scientists cultivated 2D cultures in petri dishes or 3D clusters, probing and monitoring activity from the outside.

The Princeton research took a different approach. To build the device, the team created a 3D mesh of microscopic wires and electrodes supported by a thin layer of epoxy. They then cultured tens of thousands of neurons into a vast 3D network that can perform computation, using the mesh as a scaffold.

According to the researchers, this new approach “enabled them to record and stimulate the neurons' electrical activity at a much finer scale than past approaches”. Over the course of six months, they observed how the network developed, tested techniques to reinforce or weaken links between key neurons, and eventually trained an algorithm to identify recurring pulse patterns.

To test the system, the researchers presented two distinct patterns in separate experiments, and it successfully differentiated the patterns in both cases. The team aims to progressively scale the device to perform increasingly complex tasks.

According to the paper's first author, Kumar Mritunjay, a postdoctoral researcher in electrical and computer engineering, the technology could "not only help uncover the computing secrets of the brain but can also assist in understanding and possibly treating neurological diseases.”

The original aim of the research was to investigate fundamental problems in neuroscience by studying the activities of living brain cells. That aim remains. However, the researchers realized that it could also play a role in solving one of AI’s key bottlenecks: power consumption.

“The real bottleneck for AI in the near future is energy,” said Fu. “Our brain consumes only a tiny fraction — about one millionth — of the power consumed by today’s AI systems to perform similar tasks,” said Tian-Ming Fu, assistant professor of Electrical and Computer Engineering and member of the research team.

The researchers hope the device may reveal some of the secrets behind this, making it possible to replicate the discoveries and solve AI’s power consumption problem.

The paper was published in the journal Nature Electronics.

Samsung’s production numbers reportedly plummeted significantly during a one-day strike by the company’s labor union. According to Seoul Economic Daily, citing union officials, the company’s memory fab output fell by 18% while its contract chip foundry plunged by 58.1%. These numbers only affected the night shift after the April 23 strike, with the “Joint Struggle Headquarters” saying that it will conduct a larger 18-day labor action if company management fails to reach a deal with workers. The union estimates that the walkout, which would last for more than two weeks, will cost the company KRW 30 trillion, or over $20 billion. It also threatened to mobilize personnel assigned to the fabs’ “safety protection facilities.”

The disagreement between the union and management stems from the company’s refusal to allocate 15% of its operating profit as a bonus for its workers, amounting to around $27 billion (about KRW 40 trillion), and would net chip fab workers around $400,000 each. Aside from this, the union also demanded a 7% increase in pay and a removal of the 50% bonus cap. The management made a counteroffer of a 10% operating profit bonus, 6.2% wage increase, and preferential mortgage loans, among other benefits, but it seems that this wasn’t enough for the union.

The group pointed out that SK hynix, Samsung’s biggest domestic chip rival, has given its workers a performance bonus amounting to 10% of its annual operating profit and removed the cap on the amount. This means that their bonuses only equate to less than 30% of what the SK hynix workers get, even though Samsung is 60% larger in terms of market capitalization.

Initial estimates from the police suggested that over 30,000 people attended the strike, but the union said that about 40,000 of its members were present in yesterday's action. This is a massive number and is said to represent nearly a third of the company’s semiconductor fab workforce. A general strike lasting several days would cripple operations, reducing Samsung’s advantage of being the first company to mass-produce and deliver HBM4 memory to its customers. It could also potentially exacerbate the global memory chip shortage, resulting in longer delivery times and potentially higher prices for everyone.

Unless the two sides come to an agreement, the union will kick off the general strike on May 21. Supra-Company Union Samsung Electronics chapter head Choi Seung-ho also said that he has submitted a notice to the Seoul Yongsan Police Station that the group will hold a rally in front of the residence of Samsung Chairman Lee Jae-yong on the same day.

Thousands of Samsung workers took to the streets today outside the company’s main chip fab, demanding a greater share of AI profits. According to the police, some 30,000 people attended the rally in Pyeongtaek, South Korea, more than 30 miles south of Seoul, while organizers claimed that over 39,000 were present. Bloomberg reports that the Samsung labor union wants 15% of the company’s operating profit to be shared with chip-division workers and to remove the 50% bonus cap, as well as a 7% increase in pay. If this bonus is approved, that would amount to $27 billion (more than KRW 40 trillion), averaging more than $400,000 per worker.

However, it seems that this might be a bit too much for Samsung management. It made a counteroffer of a 10% allocation of operating profit to bonuses, a 6.2% wage increase, and additional benefits like preferential mortgage loans. The union has apparently rejected this proposal, pointing out that rival SK hynix has allocated 10% of its annual operating profit for performance bonuses and also removed the maximum bonus limit for its workers. The Samsung workers argue that their bonuses only equate to less than 30% of what comparable SK hynix workers get.

While Samsung has fallen behind its homegrown rival when it came to HBM3 dominance, it has since caught up after it became the first company to deliver next-generation HBM4 chips. “The company has spoken of crisis every year,” Samsung labor union head Choi Seung-ho told the publication. “But in the midst of those crises, it was not management that sustained Samsung Electronics. It was the employees here — the union members — who made the company the world’s leading semiconductor producer, who manufactured, improved processes, worked through the night, and raised yields.”

If the two parties fail to resolve the issue, the union says it will conduct an 18-day general strike starting May 21. This massive number of workers walking out of the posts will cause massive disruption for the company, with TradingKey estimating losses of around US$20.3 billion (around KRW 20 trillion). “Samsung will continue to make efforts to reach a swift agreement in wage negotiations,” a company spokesperson told Bloomberg.

Microsoft is currently facing a lawsuit from the UK Competition Appeals Tribunal (CAT) after it was alleged that the company is charging higher wholesale prices for Windows Server for customers using Amazon AWS, Google Cloud, or Alibaba Cloud. This makes Azure, Microsoft’s cloud services provider, cheaper than the competition as its competitors have to eventually pass on the increased costs to their clients, reports Reuters. Competition lawyer Maria Luisa Stasi filed the case at the Competition Appeal Tribunal in late 2025, representing nearly 60,000 businesses that used Microsoft software on competing cloud services, with her team suggesting that the claim is worth about USD 2.8 billion (or about GBP 2.1 billion).

The company argued that Stasi did not show a concrete way of computing for any alleged losses, so the case should be thrown out. However, the London tribunal certified the case, and it’s now proceeding to trial. Microsoft is planning to appeal the decision and told the publication, “We also dispute the underlying allegations by the class representative, and today’s decision makes no final determination on those claims.”

During a previous hearing, Microsoft said that its strategy of integrating Windows Server with Azure while also licensing it to rivals will help competition. However, the British Competition and Markets Authority (CMA) discovered that this licensing practice is “materially disadvantaging AWS and Google” in mid-2025, with the government body opening another investigation into the company’s licensing practices. Other regulators in the U.S. and Europe are also looking into Microsoft and other cloud computing firms. Some sources said that the Federal Trade Commission (FTC) started scrutinizing claims late last year that the company has put punitive licensing terms in its productivity suite that make it harder for customers to switch to rival platforms from Azure.

This isn’t the only legal trouble that Microsoft faces in the UK. The company is currently appealing a ruling handed down by the same authority, which said that reselling perpetual licenses, including those of Microsoft Office and Windows, is legal and valid. This stemmed from a lawsuit by ValueLicensing, after it argued that Microsoft’s contract, which prohibited reselling of licenses issued before, is against the law. It follows a similar argument in a UsedSoft complaint filed against Redmond more than a decade ago, but Microsoft used a copyright infringement argument this time. It said that Word and other apps contained graphics that are protected by creative work, which the UK tribunal rejected.

Researchers at the University of Würzburg have demonstrated, for the first time, that chaotic growth in a 2D quantum system follows the Kardar–Parisi–Zhang (KPZ) equation, confirming a 40-year-old physics theory. For decades, physicists have believed that even highly disordered growth — from spreading flames to growing bacteria — follows hidden statistical rules.

Until now, the KPZ model, which describes how rough, uneven surfaces evolve under random conditions, had only been verified in simple, single-dimension systems, as extending it to more realistic 2D environments remained experimentally out of reach due to the extreme speeds and scales involved. The researchers’ findings, published in the Science journal, close a long-standing gap in the field, proving that the theory does indeed extend to 2D systems

This breakthrough improves scientists' understanding and modeling of complex growth processes in real-world, non-equilibrium systems.

“When surfaces grow — whether crystals, bacteria, or flame fronts — the process is always nonlinear and random. In physics, we describe such systems as being out of equilibrium,” explained Siddhartha Dam, a postdoctoral member of the research team and co-author of the paper.

“Engineering a system capable of simultaneously measuring how a non-equilibrium process evolves in space and time is extremely challenging — especially because these processes unfold on ultrashort timescales. That’s why verifying the KPZ model in two dimensions has taken so long. We have now succeeded in controlling a non-equilibrium quantum system in the laboratory — something that has only recently become technically feasible,” he continued.

To achieve this, the team engineered a highly controlled quantum system using a gallium arsenide (GaAs) semiconductor cooled to −269.15°C (−452.47°F), near absolute zero. By continuously illuminating the material with a laser, they generated short-lived hybrid particles known as polaritons — a mix of light and matter that form and decay within picoseconds.

These polaritons behave like a rapidly evolving “growth” system. As they are created and spread across the material, their distribution changes in both space and time, allowing researchers to track how the system develops under inherently random conditions.

Using spectroscopy and Michelson interferometry, the team was able to precisely monitor this evolution, capturing how fluctuations in the system scale and spread. Their analysis revealed that the behavior of the polaritons closely matches the statistical patterns predicted by the KPZ equation in two dimensions.

Nvidia has issued a statement to Tom's Hardware denying that it has been in talks for over a year to purchase a major PC manufacturer — responding to a rumor published by SemiAccurate that the AI chipmaker is close to a decision on whether the deal will push through.

The tech publication said it has spent over a year following this story, ever since it caught wind of the potential deal in late 2024. SemiAccurate also added that it was just talks, and there’s no guarantee that a transaction will even materialize. We have followed up with Nviida for a clarifying question as to whether Nvidia is in discussions with a server OEM, as mentioned in the original report.

“The media report is false; Nvidia is not engaged in discussions to acquire any PC maker,” an Nvidia spokesperson told Tom’s Hardware. Nvidia was a relatively niche computer parts manufacturer with a share price of less than a dollar until cryptocurrency mining put it in the spotlight. Everything changed when OpenAI released ChatGPT near the end of 2022, leading to a technology race between AI companies looking to create the most powerful AI model. Nvidia was perfectly positioned when this happened, and it became the biggest company selling the proverbial shovel in the AI gold rush.

This rush has allowed the company’s stock price to skyrocket, turning it from a relatively unknown GPU manufacturer to the most valuable company on Earth. The massive hoard of money that Nvidia made from selling its chips has allowed it to go on a shopping and investment spree. It has purchased AI chip startup Groq for $20 billion, took a $2 billion stake on Synopsis, sank another $4 billion into photonics firms, and invested $2 billion on Marvell to deepen their NVLink Fusion partnership — all of which has been reported in the last six months.

Aside from that, Nvidia is expected to launch its N1/N1X Arm processors for laptops in 2026, with the next-generation N2 slated to arrive next year.

J.P. Morgan said in late 2025 that the company is planning to go beyond AI GPUs and components and into complete AI servers. The rumor that Nvidia is looking to purchase a PC manufacturer seems less far-fetched when you combine this with Team Green’s push into CPUs and SoCs. The rumor resulted in a jump of more than 5% for both HP and Dell — two major PC manufacturers whose product lines cover both PCs and servers.

Nvidia has denied the story, however, saying that it did not enter into talks at all. This is probably good news for many of Nvidia’s partners, as the company building its own servers would have seen the company competing directly with its customers. And even if a deal was pushed through, it would have faced deep scrutiny by industry watchdogs and regulators.

Many news outlets are reportedly blocking Wayback Machine from archiving their pages, apparently because they fear that AI companies will abuse fair use policies and train their models on the snapshots of old articles. This risks reducing society’s collective access to historical news stories, as well as other critical information, especially in an age where misinformation is in abundance, and AI large language models (LLMs) hallucinate convincing answers. Wired reports that 23 major publications currently block ia-archiverbot, Internet Archive’s commonly used crawler, including USA Today and The New York Times. Ironically, the publication pointed out that some of these outlets use Wayback Machine in their reporting.

Many libraries and newspaper offices used to have a rich repository of archived volumes, with people accessing them to gain insights into historical records. But as the world abandoned print journalism and preferred the convenience of online newspapers, these archives are no longer updated; we must rely on online archiving services like Wayback Machine to serve as the modern historical record.

There has been some pushback from publications regarding archiving, but the legal system has established that what the Internet Archive is doing is legal and part of fair use. “Courts have long recognized it’s often impossible to build a searchable index without making copies of the underlying material,” the Electronic Frontier Foundation said. It also added, “The copying served a transformative purpose: enabling discovery, research, and new insights about creative works.”

It could be argued that the newspapers and publications themselves could handle their own archiving, but it’s in the public interest that a neutral third party handle record-keeping. After all, it’s easy to change online articles to change the record, and while many outlets are trustworthy, some are also owned by big corporations that could potentially benefit from the control of the historical narrative. Besides, it’s commonly known that outlets sometimes update articles, whether openly or in secret, so an archive like the Wayback Machine is also useful for tracking changes like these. Archive services can also be used for keeping records of publications that have since gone defunct and whose content would have been otherwise lost to history.

Companies abusing fair use policies to train AI LLMs is indeed a valid concern for both media companies and other platforms that host massive amounts of data. But preventing archiving services, such as the Wayback Machine, will do society more harm than good. Hopefully, not all is lost with archiving — Wayback Machine director Mark Graham is reportedly in talks with several outlets so that the archiver’s bot could gain access to these websites once more, while a coalition of journalists and other stakeholders have signed a letter in support of the Internet Archive and its mission of providing universal access to all knowledge.

“There is a very real scenario in which personal computing as we know it is dead,” says Framework founder Nirav Patel.

In a blog announcing the company’s Framework [Next Gen] Event 2026 on April 21, Patel decried the “winner takes all” race currently happening in the computer industry, especially as various AI tech companies are consuming memory and storage chips, and even processors, at an unprecedented pace. He noted that despite its achievements in helping push for a more repairable, upgradable, and customizable laptop ecosystem, the personal computing market is under incredible pressure.

Patel's remarks stem from the massive demand that the buildout of AI infrastructure has placed on computing and other resources. What began as a GPU shortage from 2023 to 2025 eventually turned into a memory and storage chip shortage that started in late 2025. Now, we’re seeing inklings of a CPU shortage, as data centers now demand massive amounts of server CPUs to power AI agents.

This also does not include the massive increases in electricity costs as energy suppliers and grid operators invest in new power plants and upgrade infrastructure to handle the massive amounts of power that data centers demand.

Unfortunately, these AI companies are backed by massive investments, meaning the average consumer has no recourse against these tech giants. “It’s clear that the fundamentals of computing and electronics have changed. The computer in the cloud has increasingly greater economic output than the computer in the hand.

This means that to the extent that there are constraints on the supply that feeds both, the cloud will win every time,” Patel said. He also added, “The industry is asking you to own nothing and be happy. Computers are no longer a bicycle for the mind. They are becoming the self-driving car that takes you directly to the destination.”

Still, Framework said that it will not take this lying down. Its event announcement also doubled as its own manifesto, saying that “as long as there is a person in the world who still wants to own their means of computation, we will be here to build the hardware that enables it,” and that it “will always be fighting for a future where you can own everything and be free.”

This niche company has been going against the grain of the wider laptop industry trend of non-repairable, non-upgradeable parts. Framework laptops are fully upgradeable — from memory and storage to the GPU, motherboards, and even the display.

The Framework Laptop 16 (RTX 5070) is one of its latest products, which proved that you can actually upgrade the graphics cards on a laptop; something that Dell tried (and failed at) with its Alienware Area-51m gaming laptops. The ever-increasing RAM and SSD pricing have got the entire computing industry on edge, but Framework is providing monthly updates to the community to help navigate these shortages.

A local TV channel has successfully sent a YouTube strike to every video on the Italian version of the platform that used the DLSS 5 trailer, including Nvidia itself. Gaming content creator NikTek said on their X account that the Italian media company used footage from the DLSS 5 trailer for its own coverage. However, it seems that an overzealous employee of the Italian company sent out a mass DMCA complaint, and then YouTube’s AI moderators stepped in and took down every other video that had the same content.

What’s ironic is that even Nvidia, the source of the clip in question, was taken down by YouTube. While it’s easy to blame a company or its employees for this misstep, the larger issue here is that YouTube seemingly took action without looking at the details of the complaint. The video platform says that it uses AI technology for content moderation: “In our systems, AI classifiers help detect potentially violative content at scale, and reviewers work to confirm whether content has actually crossed policy lines,” the company said on its blog. “AI is continuously increasing both the speed and accuracy of our content moderation systems.”

However, many creators are complaining about YouTube’s use of AI technology. It is estimated that the platform terminated more than 12 million channels in 2025 due to violations of its terms of service, most of which have been flagged by AI. However, some of the affected creators complained that the reasons for the takedown are inaccurate or false positives. Some even said that their appeals were rejected within a few minutes of sending them, suggesting that the case did not even go under human review.

This isn’t the first time that an original video was taken down by another channel that also used it for its own coverage, but Nvidia, by far, could be the largest victim yet. Furthermore, many other creators who used the same clip for their reaction videos have also been affected.

While Nvidia might have the muscle and resources to get YouTube to reinstate its video (which hasn’t happened yet at the time of writing), these smaller creators would likely have a harder time getting their videos back. More importantly, the takedown might add a strike to their account, which is something that many try to avoid, as it could potentially lead to the banning of their channel or account.

Oracle has reportedly begun mass layoffs based on reports by various people who worked at the company. According to the BBC, one employee noted that around 10,000 positions have been affected based on the number of active staff accounts in the firm’s Slack messaging system so far. Other people from across the company have also made posts about the situation on social media, like LinkedIn.

“Today, Oracle conducted a significant reduction in force that impacted some of the most talented, dedicated, and high-performing people I’ve had the privilege of working alongside,” Michael Shepherd, a Senior Operations Manager at Oracle, said on LinkedIn. “Let me be direct: this was not a performance action. The individuals affected were not let go because of anything they did or didn’t do. Many of them are the people you call when something is truly broken, the ones who show up early, stay late, and carry institutional knowledge that took years to build.”

The company has not released a comment about its global reduction in force at the time of writing, but there have been reports that it would axe thousands of positions and freeze hiring earlier in March. Rumors say that the cloud infrastructure firm is culling its numbers after investing billions of dollars on hardware and betting big on AI. However, this much spending means that the company is expected to be in the red until 2030.

It’s rumored that many of the affected employees will be replaced by AI, a disturbing trend after Microsoft’s AI boss said that AI can replace every white-collar job in 18 months. Oracle isn’t the only one that’s been cutting employee numbers in recent months, with Dutch chip equipment maker ASML laying off 1,700 management workers and EA reducing headcount in Battlefield-related studios. Meta has not laid off employees yet, but there have been rumors that the company is reducing employee bonuses for the second time in a row as it spends billions on AI hardware and talent.

This isn’t the only problem that Oracle is currently facing. Earlier this year, some of its bondholders sued the company for misleading statements after it borrowed $38 billion two months after releasing $18 billion of notes and bonds in September 2025. There have also been reports of disagreements between OpenAI, Oracle, and SoftBank regarding who should have the ultimate control of the Stargate project, with the planned 600MW expansion at the Abilene, Texas, campus getting scrapped.

These AI-driven layoffs are sounding disturbing for the average worker, especially as many industry experts and company leaders have been warning about the disruption that AI will bring to the job market. However, OpenAI boss Sam Altman said that companies are just using 'AI washing' as an excuse for these job cuts that they would have otherwise done anyway. But whether this is true or not, it gives little comfort to the affected Oracle employees, especially during these uncertain times.

Mimms Museum of Technology and Art in Roswell, GA, will open the doors to its iNSPIRE: 50 Years of Innovation from Apple exhibition on Wednesday, April 1. The exhibit organizers say this is a first-of-its-kind exhibition and will feature more than 2,000 artifacts across 20,000 square feet. That makes it “the World’s Largest Collection of Apple Products,” according to the Mimms PR department. We’ve got some pre-opening images and details to share.

According to our research, Mimms won’t actually be showing off the world’s largest Apple hardware exhibit, though. That prize goes to the All About Apple Museum (Savona, Italy) with over 9,000 items, and a living museum area where visitors can interact with many old computers and devices. Mimms comes in second place, though!

A ribbon-cutting grand opening event for iNSPIRE is scheduled for 10am on Wednesday, with public access to the extensive artifacts and installations beginning at 12 noon.

Yes, it really has been 50 years since Apple Inc. was founded by Steve Jobs, Steve Wozniak, and Ronald Wayne. Its 1976 launch product was the Apple I personal computer. A year later, the company would be incorporated by Jobs and Wozniak as Apple Computer Inc.

To mark this significant historical milestone, Apple will start its half-century celebrations in earnest. It even has a new Instagram account lined up for the event. However, Apple fans who’d prefer something more tangible and physical to get into a celebratory mood can visit Mimms in Roswell.

At the iNSPIRE: 50 Years of Innovation from Apple exhibition

Central to the exhibit are the 2,000+ artifacts from what has become one of the world’s most influential brands. There will be oodles of hardware to see. Besides that, visitors can check out “Rare collector’s items, unique documents, and behind-the-scenes stories of Apple’s groundbreaking product.”

As a technology and art museum, Mimms isn’t just shoving a load of old tech in glass boxes, of course. The museum is promising some hands-on displays of Apple’s early computers and devices. Moreover, “immersive, creative installations inspired by Apple’s most memorable campaigns,” will be part of the visitor experience.

Spoiler alert – we have some details and photos of some of the immersive exhibits, like the Motion-based trivia experience, iCloud floor interactivity, and the iPad drawing wall.

Museum admission is $22 for adults, $16 for kids (4 to 17), and a family of four can enter for $66. Mimms also offers concessions to seniors, military, and college students. Taking photos and sharing them online is encouraged.

Coincidentally, the Apple Museum Utrecht (Netherlands) will open its doors for the first time on April 2. There, you will be able to visit the massive wall of 100 iMac G3s, and enjoy walking around a full-scale replica of the Jobs' family garage.

The U.S. Supreme Court has ruled in favor of Cox Communications after Sony Music Entertainment and several other labels sued the company for copyright infringement back in 2018. The internet service provider (ISP) actually lost the case back then after the music labels said that the company did not terminate its subscribers despite being repeatedly flagged for downloading and sharing pirated music. This ended with the jury awarding the plaintiffs $1 billion in statutory damages, although this was overturned on appeal, according to Engadget.

“Countless people use the Internet for legal activities, but some use it to illegally share copyrighted works, such as songs and movies…In this case, however, instead of suing those infringers, the copyright owners sued petitioners, Cox Communications, Inc., and its subsidiary, who provided the internet connections that the infringers used,” Justice Clarence Thomas wrote in the decision. “Under our precedents, a company is not liable as a copyright infringer for merely providing a service to the general public with knowledge that it will used by some to infringe copyrights.”

This ruling sets a precedent, saying that ISPs are not liable for the actions of their users. Although Cox has a clause in its service contract that prohibits users from distributing pirated content, it seems that quite a few of its six million subscribers ignore it. A Supreme Court of the United States document [PDF] showed that MarkMonitor, a firm that tracks the illegal uploading and downloading of pirated media, sent the ISP 163,148 notices that identified infringing IP addresses over a two-year period. But, out of its nearly 6 million subscribers, Sony says that the company only terminated 32 subscribers for their alleged illegal activities.

The court says that Cox has no contributory liability to the copyright infringement cases that Sony brought against it, as it needed to prove that the ISP either “affirmatively induced the infringement” or that it “sold a service tailored to infringement.” Furthermore, the court pointed out that “Internet service providers like Cox have limited knowledge about how their services are used; they know which IP address corresponds to which subscriber account but cannot distinguish individual users or directly control how services are used.”

British Army veteran Jon L. Noble has shared a heartwarming update on his first 100 days with a Neuralink brain computer interface (BCI) implant. Noble variously describes the journey from day 0 to the present as “science fiction… magic… brilliant… addictive… overwhelming and incredibly motivating.” Along the way, he’s moved from Mac newbie to power user. As for gaming, “I’m now raiding, and exploring Azeroth hands-free at full speed — no mouse, no keyboard, just intention,” enthuses the World of Warcraft fan, also known as P-18 (patient number 18).

For some background to Noble’s situation, the Financial Express reports that the ex-paratrooper was paralyzed from the shoulders down after a driving accident in 2016. He volunteered for Neuralink clinical trials and got the Neuralink N1 implant in London in December last year.

Noble’s first 100 days as an N1 implant patient have whizzed by, and he starts his recollective Tweet by summing up, “Looking back, the whole journey feels like science fiction that somehow became my everyday reality.”

On day zero, he went in for surgery, with a robotic system inserting 1,024 threads into his motor cortex. He was well enough - “alert and in good spirits” - to go home the next day. By day seven, the scar was already fading, and Noble says he “felt sharper and more positive than I had been in years.”

The fun started in week two, when the N1 was paired with a new Apple MacBook, says the ex-paratrooper. Neuralink engineers helped him get used to flexing his new BCI, and by week three, Noble says “scrolling, clicking, typing — all mind-controlled,” was second nature on the Mac. “I went from total Mac newbie to power-user faster than I ever expected,” he joked.

Riding into the realms of Azeroth

There’s a bit of a gap in the timeline shared, but by day 80, Patient 18 recalls he was “ready for the big leagues.” Of course, gaming presents quite a different experience to general computer interaction, with the stress on timing, accuracy, and not forgetting strategy. Noble admits he started off feeling clunky.

“Once my brain and the BCI synced, it was pure magic,” explained the ex-paratrooper. “I’m now raiding, and exploring Azeroth hands-free at full speed — no mouse, no keyboard, just intention. It’s honestly brilliant. The freedom is addictive.” You can see a video of Noble playing the game if you expand the embedded post above.

Noble also takes the opportunity to thank all the positive folks he’s had the pleasure to interact with on social media since his implant. It is good to hear of the overwhelming positivity and excitement communicated in the thousands of messages he has received.

With these first 100 days behind him, Noble relishes what the next 100 will bring. “The N1 didn’t just give me a new way to use a computer — it gave me a new way to live,” seems like a good summary of the life-changing experience gained from a capable BCI.

Hardware and software laden with ads have, unfortunately, become part and parcel of modern life, but there are occasions when the hunt for revenue goes too far. One of those cases comes from Hisense, known across Western markets as a budget electronics brand. The firm's TV sets have repeatedly come under fire for forcing non-skippable ads when switching inputs, turning the TV on, navigating to the home screen, and even when switching channels — all changes that took effect unilaterally after purchase, reportedly even for users who had all ad-related options disabled.

The affected models are mostly but not exclusively lower-end units with Hisense's VIDAA operating system, recently rebranded as Home OS. The vast majority of reports come from Hisense TV owners, but we saw at least one such complaint about a Toshiba set. The operating system is also licensed by Schneider, Akai, and Loewe, among multiple other brands.

This issue came to light recently due to press coverage, but it dates back at least a year, and possibly three, depending on how you count. The earliest notable report dates to 2022, when a user spotted an ad option in their input selection menu. These complaints have gotten more frequent with time, with some people noticing they were forced to watch ads when they turned on their sets. Reports from the last two weeks display the more aggressive tactic of forcing ads when changing TV inputs. Spanish outlets El Español and La Razón covered reports that users were being delivered ads when simply changing channels, too

The situation gets sketchier when reading through user discussions of how to avoid this madness-making behavior. Most suggestions for avoiding the ads are predictable, such as changing the TV's DNS servers or disconnecting it from the internet entirely. Still, a common solution is to contact Hisense support with the TV's unique ID at the seemingly Australian address service.tv.au@hisense.com.

Users who contacted support via email reported that the ads were disabled on their sets, which raises the question of whether Hisense is simply managing delivery on the ad server side or has deeper access to the TVs in question. Additionally, while a sufficiently motivated or technically minded user will forge all the way through this route, it's reasonable to expect that the public at large would grudgingly accept the ads if the sets are outside their store's return window.

Then there's the matter of the location of these incidents. Most reports seem to come from British and Spanish users, but we also found a German-language post and screenshots of a TV set in German. La Razón dug into this matter and published a statement from Hisense that arguably raises more questions than it answers.

Hisense says the ads did not stop owners from "using their devices normally" (a fact reiterated three times) and that the ads were part of "spot tests within the Spanish market," meant to "evaluate certain advertising formats linked to free content within the platform itself."

Given that user reports span multiple countries and a wide time frame, and that Hisense has an Australian email address that answers customer queries on the subject, the situation looks as clear-cut as the circular economy around AI investments.

As far as we can tell, the list of countries from which complaints originate is part of a list of nations covered by an advertising agreement between VIDAA and Teads. The entire statement from Hisense follows; note the translation is our own.

"Regarding the recently-published information about the alleged inclusion of mandatory advertising in Hisense television sets, the company wishes to clarify that in no circumstance did its devices force users to watch ads to use them normally.

The aforementioned situation is exclusive to a spot test performed in the Spanish market within the scope of the VIDAA platform, the television sets' operating system. This test's objective was to evaluate certain advertising formats linked to free content within the platform itself.

In no circumstance did the test affect the standard functionality of the device nor did it limit access to its main features. The users could and can continue to normally use all HDMI inputs, external devices, consoles, subscription streaming apps, or standard broadcasts without any type of interruption or obligation to watch advertisements.

This was a temporary and finalized market test. The aforementioned advertising format has now been removed from Spain. Hisense maintains its commitment to a quality, transparent user experience, based on freedom of choice, guaranteeing that the usage of the television set and its main features are not conditional on watching advertisements."

While Hisense's statement suggests this "test" was regionally limited, the ongoing complaints about intrusive ads on its TVs suggest that the presence or absence of those ads is more widespread and longer-lasting than claimed. In any event, users seeking a less burdensome TV-watching experience might want to steer clear of Hisense's hardware, as they risk unpleasant surprises when performing basic tasks on their sets.

The biggest AI tech companies, including Amazon, Meta, Microsoft, and Google, are expected to join President Donald Trump at the White House today to sign the new “ratepayer protection pledge” in a bid to keep electricity prices under control. Trump said during the State of the Union address, “We’re telling the major tech companies that they have the obligation to provide for their own power need.” According to Bloomberg, Meta, Amazon, and OpenAI have confirmed attendance at the White House event on Wednesday, while Google declined to comment.

The massive power demands of the AI data centers have upended the electricity market, causing electricity costs for the average American to increase sharply. Some states have reported an increase of up to 36%, with wholesale power prices increasing by 267% in just five years. This is putting undue stress on the ordinary consumer, so much so that both sides of the aisle are demanding action from tech giants.

Microsoft was the first to respond to Trump’s call for AI data centers to “pay their own way” when it comes to their electricity consumption, promising to be a “good neighbor” in the communities where it is present. OpenAI soon followed suit, saying that it will fund grid upgrades and apply flexible loads to reduce stress on the grid, while Anthropic said that it will pay 100% of its grid infrastructure costs and even produce new power sources. Now, it seems that the administration wants to formalize these promises.

Energy Secretary Chris Wright reportedly said in a statement that the pledge "will deliver more affordable, reliable, and secure energy for the American people and help stop the rising electricity prices that started during the previous administration," while ensuring the U.S. wins the AI race.

However, experts are skeptical whether these promises and pledges will hold water in the long run. “The ratepayer protection plan is a show designed to sweep this issue under the rug and show the White House has solved the problem,” Director Ari Peskoe of the Harvard Law School Electricity Law Initiative told the publication. “The White House has no real authority here aside from the bully pulpit.” It’s not until we see big tech companies sign separate rate structures with electricity providers, grid operators, and state governments that we will know whether these pledges are effective.

At yesterday’s State of the Union address, President Trump brought up the issue of surging power costs driven by hyper-scale AI buildouts — and proposed a solution. Announcing a new “ratepayer protection pledge,” he said companies would now be required to build their own power plants for data centers, generating and supplying their own electricity for AI workloads.

For the past few years, Big Tech has relied on thirsty data centers to fuel the AI boom, building massive sites running thousands of GPUs at once. These chips not only require energy themselves but also need to be kept cool, which adds to the overall power needs. So far, these companies have just been just plugging into the grid and buy electricity conventionally, but this has stretched the grid thin.

Now, for everyone else in the area, power has become more expensive because their locality is pulling harder from the network. "We have an old grid. It could never handle the kind of numbers, the amount of electricity that's needed," said Trump. Last year, a report claimed that energy prices have already risen by up to 36% in some states, while another pointed toward the situation getting worse.

Data centers are said to account for 12% of the total power in the national grid by 2028, up from just 4% back in 2018. This drastically affects the average person, who is forced to pay more for the same household power usage and whose backyard is now a perpetually-humming living being — while the corporations actually behind this surge remain much more shielded against the price hikes.

To combat this, Washington's new plan is to ask these companies to fuel their AI ambitions by themselves. White House spokesperson Taylor Rogers told CNBC that "under this bold initiative, these massive companies will build, bring, or buy their own power supply for new AI data centers, ensuring that Americans’ electricity bills will not increase as demand grows."

This pledge will reportedly be signed by Amazon, Google, Meta, Microsoft, xAI, Oracle, OpenAI, and others when they visit the President in early March. Trump already unveiled the idea last month via a Truth Social post, where he ensured that Americans won't have to "pick up the tab" for data center buildouts.

The Trump administration has been very aggressive on the AI front in order to deter China from gaining an upper hand. The two countries were engaged in a deadlock for most of past year over the latest AI GPUs, before things cooled down with a temporary trade truce. Hence, the government is maintaining a delicate balance between encouraging hyperscalers to... scale while simultaneously keeping them in check.

"I'm telling them, they can build their own plant. They're going to produce their own electricity. It will ensure the company's ability to get electricity, while at the same time, lowering prices of electricity for you," said Trump at the SOTU address. Now, it's only a matter of time before this "ratepayer protection pledge" is formalized, but broader reaction from concerned critics will be predicated on actual, real-world impact.

What do airlines, banks, and insurance companies have in common? Besides being an absolute pain to deal with, they all rely on COBOL and IBM mainframe computers as core infrastructure. The computing giant's stranglehold over those markets may finally begin to crack, though. Anthropic has announced COBOL-specific functionality for its Claude AI bot, and IBM's investors responded with a resounding 13% drop in stock prices.

Anthropic published its ideas in a blog post, and the company seems to know its target market quite well. There's a Code Modernization Playbook available for download, and existing YouTube videos letting Claude Code loose on a COBOL illustrate the concept. Mixing "COBOL," "AI bot," and "YouTube" in the same sentence is utterly anathema to common logic... and yet here we are.

For practical purposes, COBOL only runs on one type of system, supported by one set of people: IBM's mainframes and its engineers. That means the company has enjoyed multiple decades of telling clients how many zeros the gigantic bills will have tacked on for the next period. The unforgiving nature of that stranglehold means that any attempt at breaking it is most welcome by its existing customers, and a serious threat to IBM as a business.

If you've ever interacted with social security, public administration, healthcare, government, finance, insurance, automotive, retail, or airlines, you've touched a COBOL system at some point in your transaction, even if it was 30 layers deep. Similar to gravity, the language is invisible and yet affects every part of the modern world.

A cynical view of the situation would say that the systems running COBOL are meant to be 100% accurate 100% of the time, a notion that doesn't lend itself very well to the "probabilistically correct" that LLMs can offer. Even still, as I've attested myself, a good bot is a power multiplier in the hands of a competent developer, and can also lower the barrier of entry for young folks trying their hand at wrangling old systems.

The language harkens back to the 1960s, proposing itself as a human-readable language targeted at business transactions, using full decimal-point math as default, in contrast to the default floating-point math in other languages. True to its proposal, it revolutionized business computing, becoming entrenched across almost every sector of note, and was never truly replaced.

The situation doesn't just revolve around IBM's monopoly, though. Most well-versed COBOL programmers are retiring and dying, making their skills rarer and more expensive. The COBOL systems invariably run business-critical operations that cannot afford any downtime whatsoever, and are chock-full of proprietary data formats and business logic that is not documented, and understood only by a few greybeards — if at all.

If you're wondering why COBOL just wasn't up and replaced with something else, know that any rewrite attempt must (a) reverse-engineer miles-long business logic; (b) reverse-engineer the underlying data structures; (c) reimplement said logic and structures while being careful to always use fixed-point decimal math; and (d) execute a perfect transition with minimal to zero downtime.

Even when all those conditions can be true, COBOL systems are often so interconnected that it's unfeasible to replace just the one, as is the case with airlines. And heaven help you if you're in the financial sector, as you'll have to undergo extremely long-winded tests and audits, adding months to any deployment.

There's a well-known joke among programmers that almost certainly originated from COBOL: "When I wrote this code, only God and I understood what it does. Now... only God knows."

Data center buildouts have regularly been in the news for their massive power needs, the strain they put on the grid, and the energy prices that can increase for residents in the general vicinity. The buildouts usually require years of planning and approvals for grid connections. Those timeframes are a preposterous notion for companies invested in the AI gold rush, who have started sidestepping the issue entirely by building their own power generation, fueled mostly by natural gas turbines.

A cursory look at the situation suggests it's a win-win for everyone, as it avoids stressing local grids and accelerates buildouts. However, the quick-patch solution hides many pitfalls, including, but not limited to, carbon emissions and local pollution. This "shadow grid," as the Washington Post (WaPo) calls it, reportedly includes at least 47 data center buildouts across the nation, according to a report by energy-tracking company Cleanview.

Many of these data center builds have colossal power requirements. The off-grid GW Ranch project in West Texas will reportedly use as much power as Chicago, combining natural gas and solar panels to feed itself. A similar story is unfolding in Mason County (West Virginia), where the expected gas demand is enough to serve 1.5 million residences. Usage could eventually quadruple, enough to serve every West Virginia home. Cleanview's Michael Thomas described the situation as "catastrophic for climate goals."

Whereas popular opposition to the construction of new data centers has grown fierce, WaPo says that some states have passed laws enabling off-grid data centers by tweaking rules on where power plants can be built and who can build them. The unexpected appearance of the plants has left local residents frustrated by the lack of transparency and agency, especially as most companies behind the buildouts reportedly try to keep their names off the projects.

Even without accounting for environmental concerns, data centers may still impose indirect costs on local utilities, as AI's money lets builders outbid utilities for equipment, leaving local grids to bear higher maintenance and expansion costs. Indeed, gas turbines are already sold out through 2030, surprising no one.

To their credit, AI companies are investing in nuclear SMRs and fusion research, but those take time, while the AI gold rush demands immediate rollouts of data centers.

Critics also point out that the gas plan may not be as solid as the entrepreneurs think, given that gas turbines cannot operate around the clock like data centers do, apparently spending a third of their time offline. Adding to that, data center builds are "trying to rush to market with a bunch of clankety old stuff that was headed to the scrapyard, or with dozens to hundreds of small generating units strung together," according to Eolian's CEO.

AMD's AI chips will be used as debt collateral in a deal to financially support a cloud startup buying the company's processors, The Information reports. It's the first known occurrence of the practice with AMD's chips, creating a circular deal similar to those that Nvidia has been making to increase GPU sales.

The $300 million loan from Goldman Sachs to the cloud startup Crusoe will be used to buy AMD's AI chips (the reporting doesn't specify which chip will be used, though presumably it will be a member of the Instinct MI450 series). Those chips will be used in a datacenter in Ohio.

AMD is set to effectively serve as guarantor for the deal, according to the report, by agreeing to rent the chips from Crusoe if the company can't find other willing customers, sources close to the deal told The Information. In the meantime, AMD will get to say it has $300 million in AI chip sales.

Crusoe is a cloud startup that claims to build data centers with cleaner energy than other companies, with more due diligence on "sites with clean, scalable power already in place."

This is far from AMD's first major AI deal. Last year, OpenAI secured up to 6 gigawatts of AMD GPU compute in a deal that could see the ChatGPT creator take a significant stake in AMD. In December, AMD and HPE entered into an agreement that will bring the AMD Helios rack-scale AI architecture into HPE’s product portfolio.

AMD is following Nvidia in using chips as collateral. Nvidia famously put up H100 GPUs as collateral to back CoreWeave's $2.3 billion loan by Magnetar Capital and Blackstone. These circular deals let the companies make big claims about how many chips they've sold and where they're deployed without startups seeing much of the actual risk.

AMD has denied reports that its next-generation MI455X accelerators may be delayed in production. Those are set for the second half of 2026. We've reached out to AMD for comment.

The U.S. Department of Commerce is reportedly planning tariff carve-outs for AI hyperscalers within the country, including Amazon, Google, and Microsoft, among others. According to the Financial Times, the move will allow these companies to acquire the chips they need to remain competitive in the AI race, while still pushing chip makers like Taiwan Semiconductor Manufacturing Company (TSMC) to continue investing in the U.S.

Taiwan and the U.S. agreed on a trade deal last month, with Washington cutting tariffs on the island from 20% to 15% in exchange for a $250-billion commitment on direct investments, plus another $250 billion in credit guarantees from Taipei to allow its companies to invest in the American semiconductor industry. The 15% tariff is set to include chips made on the island, but Taiwanese companies building fabs in the U.S. will get an exemption.

These chip manufacturers can import two and a half times their fab’s planned capacity tariff-free while their plants are under construction. That number would then fall to just 1.5 times once they start churning out semiconductors locally. TSMC is expected to allocate its exemptions to its Big Tech clients that are ordering huge numbers of AI chips, but the total quantity of tariff-free chips will still depend on the company’s investments in the U.S.

At the moment, TSMC has already committed to invest $165 billion in the U.S., with one fab already operational in Arizona. Despite this, a source told the Financial Times that U.S. President Donald Trump is delaying the signing of the agreement. “We’re going to be monitoring what unfolds after this is unveiled like hawks to make sure that the integrity of what we’re trying to accomplish with the tariffs and the rebates isn’t undermined and that this doesn’t end up being a giveaway to TSMC,” the administration official said.

On the other hand, the TSMC board recently announced a $45 billion spending package on new fabs. This is a massive change for the company, which usually spreads out its capital appropriation throughout the year. This recent move instead saw a huge chunk of the chip maker’s planned $52 to $56-billion budget for capital expenditure allocated in the first quarter of 2026. It’s unclear how much of this will be spent on American fabs, though. Although Taiwan rejected the possibility of moving 40% of its semiconductor capacity inside the U.S.’s borders, it said that the production capacity of both local and American chip production is expected to scale hand-in-hand as the AI infrastructure build-out continues, driving record demand for the latest chips.

Big spending in AI-related investments has become the new normal to the point that it's now background noise. Even still, occasionally there's a sonic boom. Just yesterday, Amazon announced that it would be spending $200 billion in 2026, or $50 billion more than predicted. Investors didn't like that, and the company's shares took a steep 9% nosedive, taking some of its friends along for the ride for a combined sell-off approaching $1 trillion.

According to the Financial Times, the Big Tech players are set to spend a $660 billion on AI investments, an amount larger than the GDP of Israel. Investors who were once very bullish on the AI race, not wanting to be left out, are reportedly starting to get cold feet.

Revenue large enough to outstrip AI investments could be looking more like a mirage than an oasis, with analyst Dec Mullarkey stating that the announced spend is "not welcome news for investors that are already fixated on when AI-related revenue will start to show up."

Amazon took the brunt of the hit, as, along with the gigantic increase in capital expenditures, investors are seemingly frowning at the possibility of the outfit cannibalizing its lead in cloud services and even retail presence for the sake of AI. Stressing that particular point, analyst firm D.A. Davidson downgraded Amazon's rating from "buy" to "neutral".

The big loser group includes Meta and Alphabet (Google), which saw their shares take around a -2% and a -3% spill respectively, for the same base reasons. Even Google's record earnings and a contract with Apple for providing Cupertino's AI services didn't help it escape investor wrath. Analyst Mamta Valechha points out that alongside key fears, investors are not appreciating the companies' lack of visibility into exactly how these investments are expected to play out.

One week doesn't make for deep financial analysis, but it's worth noting that since Monday, today's sell-off puts Amazon at around -11.3%, Alphabet at -3.15%, Meta at -7.4%, Microsoft at -7.7%, and Oracle at -9.2%.

Meanwhile, Tim Cook is probably chuckling and eating popcorn. Only a scant few months ago, Apple was strongly criticized for dropping out of the AI race and for the lack of its much-touted Apple Intelligence in its latest software releases.

The firm ultimately threw in the towel and hired Google's Gemini for that duty, but not having spent untold billions resulted in investors sending the stock price up 7.5% over the week, helped by "staggering" demand for the latest iPhones.

When Sir Tim Berners-Lee first came up with the idea for the World Wide Web back in 1989, his vision for it could be likened to a digital Library of Alexandria, with all human knowledge centralized, a cooperative attitude, and most of all, free. Fast-forward about 37 years, and while the Web did indeed bring the world together, it's fair to say that it didn't do so quite in the manner that Berners-Lee hoped for.

In an interview with the Guardian, Berners-Lee discusses where exactly things went wrong, and how they can perhaps be fixed in what he describes as "a battle for the soul of the web," noting that "it's not too late." The Web had a fairly peaceful early start without ads, heavy commercialization, and cleanly served the purposes of information and some entertainment, but started shifting around the late '90s with the dot-com boom and "charlatans", in Berners-Lee's view.

It wasn't until the polarization of the 2016 U.S. election that he had enough with the Web's toxicity, something that reportedly left him "devastated." He acknowledges that social media does not represent the entire web, but that "the problem is that people spend a lot of time on [social media websites] because they’re addictive," having later described them as "optimized for nastiness".

Interestingly, Berners-Lee notes that while early viewpoints on the internet were neutral and moral judgments were about how people used it, nowadays "the way you design a website, like Reddit or Pinterest or Snapchat, can be explicitly good," with the opposite, as now proven, equally possible.

He's since created and is promoting the Solid project, that can very broadly be described as a user-controlled, decentralized API for personal data sharing, accessed via standard web protocols. Users create their own "pods" where all their personal info lies, along with everything they publish online, and data collected about them. They can then choose to grant entities specific control over what data is accessed.

Berners-Lee believes this can bring sovereignty of personal data back in people's hands, and even went as far as raising $30 million for Inrupt, a company with a mission to "provide commercial energy and an ecosystem to help protect the integrity and quality of the new web built on Solid". Notably, though, there's nothing forcing the current web's gatekeepers like Google and Meta to use the Solid protocol — much less to copy any data they may be granted access to.

Still, he think that "the existing systems will fade to a certain extent, because people will get more excited in new systems." In this case, he's referring to developers engaging with the Solid project, who "they start coding just because of what they can imagine."

You can click through to the Guardian to read the entire interview and read Berners-Lee's thoughts on social media ban laws, the AI takeover, and other topics.

Another day, another deal earmarking large amounts of money from a big corporation to one of its customers in the AI world. The latest exchange is between Nvidia and cloud datacenter makers CoreWeave, where Jensen Huang's outfit bought a helping of CoreWeave Class A shares, for $87.20 a piece.

Before the deal, Nvidia owned just over 6% of CoreWeave, a slice that ought to have increased today to around 9%. Although those figures and today's purchase relate to standard Class A shares, some outfits are reporting that the chipmaker also owns some Class B stock — shares with 10x voting rights, privately traded.

For Huang, this investment reflects "confidence in their growth and confidence in CoreWeave’s management and confidence in their business model" [sic]. CoreWeave is one of the many companies rapidly burning through far more money than it's getting revenue, a fact that previously had some investors skittish, particularly after December when the firm revealed a plan to raise $2 billion by issuing debt to be exchanged for shares.

After today's news, though, CoreWeave's shares rallied to a 9% bump, reflecting the sizable cash injection and perhaps also Huang's outlook on the matter. Additionally, CoreWeave is apparently the first Nvidia customer getting access to its Vera CPU chip as a standalone unit, something that might have helped with the stock rise.

The green team's fresh Arm-based design was previously only available as part of an entire system board, but it's now going to be available as a standalone product, though seemingly for datacenter customers only for the time being.

As a refresher, Vera's specs are 88 cores and 176 threads, an Arm v9.2-A instruction set, 2 MB of L2 cache on each core, and 162 MB of shared L3 cache. Among many other interesting bits of info, Nvidia's Spatial Multithreading should allow each Vera core to effectively run two hardware threads, by way of divvying up resources by partition instead of time slicing them like standard SMT.

Other interesting figures include up to 1.5 TB of precious RAM per CPU, capable of pushing data at up to 1.2 TB per second. The onboard NVlink interconnect can also handle 1.8 TB/s of bytes flying by. As a rather obvious statement, the chip's super-wide design makes it perfect for AI workloads.

The FBI served a search warrant to Microsoft in early 2025 to recover the encryption keys on three laptops, with the company complying to allow the agency to access the data on the devices that it otherwise would have been unable to read. According to Forbes, Microsoft spokesperson Charles Chamberlayne said that it receives around 20 requests for BitLocker keys annually, but most of them fail because the user did not store their recovery key in the cloud.

Although there have been many requests through the years, with one Microsoft engineer even claiming that the U.S. government approached him way back in 2013 to install a backdoor in the encryption system (which he declined), this is the first recorded instance where the tech company complied and resulted in a breakthrough for the government.

BitLocker is Microsoft’s built-in drive encryption system to protect Windows 11 users. And while it’s mostly designed to prevent unauthorized access to a drive’s contents, it has also experienced some bugs that can cause a significant loss of data, especially if you forget your encryption key. Nevertheless, Microsoft backs up your BitLocker keys online by default, making it more convenient for users to unlock their drives for situations like this. However, this also makes them vulnerable to valid government requests — not just from the U.S., but from other governments, as well, with less than stellar reputations, especially in human rights.

“While key recovery offers convenience, it also carries a risk of unwanted access,” Chamberlayne told Forbes. “So, Microsoft believes customers are in the best position to decide… how to manage their keys.” Americal Civil Liberties Union (ACLU) surveillance and cybersecurity counsel Jennifer Granick also said, “Remote storage of decryption keys can be quite dangerous.”

Apple offers a similar encryption system to Microsoft’s BitLocker with FileVault and Passwords, while Meta also keeps encrypted backups of WhatsApp data. Both companies allow users to keep backup keys for these systems online, but they’re also kept in an encrypted file. So, even if a government agency requests a copy of the stored key, neither Apple nor the concerned agency can unlock it without the proper key. Furthermore, Forbes notes that neither Apple nor Meta is known to have acquiesced to a request for an encryption key.

Given how quickly the evolution of AI has upended technology across the globe and is affecting various markets, it's nigh impossible to accurately predict where anything might be headed. There's no shortage of predictions, ranging from utopia to ultimate doom for established industries. An NYT columnist, however, has one specific bet: OpenAI will be destitute in 18 months in the wake of its AI endeavors.

According to an external report last year, OpenAI was projected to burn through $8 billion in 2025, rising to $40 billion in 2028. Given that the company reportedly predicts profitability by 2030, it's not hard to do the math.

Altman's venture projects spending $1.4 trillion on datacenters. As Sebastian Mallaby, an economist at the Council on Foreign Relations, notes, even if OpenAI rethinks those limerence-influenced promises and "pays for others with its overvalued shares", there's still a financial chasm to cross. Mallaby isn't the only one thinking along these lines, as Bain & Company reported last year that, even with the best outlook, there's at least a $800 billion black hole in the industry.

The financial guru contextualizes the situation adeptly, broadly stating that it's not a matter of whether end-user AI will become technologically entrenched, but rather whether the economics of developing it will make sense in the mid- to long-term.

The analyst points out that in theory, investors would "bridge the gap between the emergence of a great technology and eventual profits", except that many AI companies seem to be burning cash far faster than they can generate income. Mallaby remarks that the newcomers are in a much worse position than "legacy" companies like Microsoft or Meta, given that the old-timers already had money-making businesses before AI came about and can (literally) afford to wait out the necessary period until the clankers deliver the fruits.

According to him, the majority of people are using free services and have no qualms switching to a competitor once their usual bot adds ads or usage limits, a fact corroborated by the myriad options available right now for all sorts of tasks.

He sees this as a temporary problem for AI providers, though, as agentic AI becomes more entrenched in daily people's lives, it'll become harder to switch, as the bots should eventually have all your shopping preferences, aspirations, and emotional profile mapped out —perhaps even better than yourself.

Mallaby does praise OpenAI's CEO Sam Altman's dollar-attracting gravitational field that raised $40 billion in investment, an amount bigger than any other private funding round in history — even more than Saudi Aramco's $30 billion. The difference is that Aramco, along with other IPO'd enterprises, had a business model and profitability, neither of which OpenAI currently enjoys.

The AI financial ouroboros certainly looks poised to eat its own tail, but there's an argument that the ophidian will only lose its newer part. There would be some irony in the AI market losing one or more of the players that started it all.

Pioneering search engine AltaVista opened its service to the public 30 years ago. The original fast and clean internet search destination launched on December 15, 1995, with an enormous (for the time) 16 million pages indexed. Within a year of its establishment, it scaled from a day-one workload of handling 300,000 user queries to tens of millions of requests every day. That’s impressive, given that the service went live ostensibly as a tech demo for DEC’s Alpha server hardware.

AltaVista rises - propelled by powerful hardware

A big part of the initial success story of AltaVista was indeed due to the hardware that powered it. As a showcase project created by Digital Equipment Corporation's Network Systems Laboratory and Western Research Laboratory, it isn’t surprising that this search engine’s performance was propped up by some of the most powerful servers of the era.

Specifically, AltaVista was launched as an Internet search service powered by a DEC Alpha 8400 Turbolaser system, say some sources. We think credit to that precise server model might be uncertain, due to upgrades over the years and exact specs being lost in the sands of time. The 8400 servers came with up to 14 CPUs running at up to 612 MHz, up to 28GB of RAM, and had 144 PCI slots.

We’ve already sketched out the explosive but almost accidental success that the youthful AltaVista would see. Its architecture, mixing a web crawler dubbed Scooter and a back-end index server called TurboVista, delivered fast and accurate results in the early days of the public adoption of the World Wide Web.

At its inception, AltaVista stood apart from the dominant Internet directories, which had become the default home pages of many. Instead of some kind of curated directory, surfers could tap into the quickly expanding full extent of the Internet thanks to Scooter and TurboVista tirelessly beavering away.

Illustrative of its rapid climb to being a top-tier Internet company, in 1996, AltaVista became the exclusive provider of search results for Yahoo! That’s the year after it launched.

Google arrives with its PageRank algorithm

Describing the rise of Google is another topic in its own right. However, to cut a long story short, its origins are found in Larry Page and Sergey Brin’s BackRub Stanford research project. That’s a good project name, as it almost encapsulates an algorithm that became the jewel in Google’s crown – PageRank.

With PageRank at its heart, Google was founded in 1998 and enjoyed explosive early growth in 1999. People liked the AltaVista-like, simple, clean search UI Google offered, and generally warmed to its weighted, relevant results, served by the PageRank algorithm’s analysis of over 50 million pages in 1999.

In 2000, Google made its business breakthrough, becoming the default search engine for one of the earliest Internet icons, Yahoo!

AltaVista falls - Changes of ownership, and years of decline

If you’ve read all the above, you can see a crossover or collision occur, as AltaVista declines and Google rises. AltaVista sought to fend off Google in the early 2000s with interface and indexing relevance tweaks. However, between 1999 and 2001, Google stole the momentum, knocking the wind out of AltaVista’s sails. An oft-quoted stat puts AltaVista vs Google user share at 17% vs 7% in the year 2000, but the tables were turned very shortly after that.

Famously exerting further downward pressure on AltaVista were a string of acquisitions, where it changed hands from DEC, to Compaq, to CMGI, to Overture Services, and then by acquisition of Overture to Yahoo! Along the way, AltaVista moved away from the clean search results first UI into becoming an all-singing all-dancing portal, helping nail its coffin shut (Hi Google).

Though it was formally shut down by Yahoo! in 2013, nearly 18 years after it first made WWW waves, many of us remember AltaVista fondly. It pioneered the fast and uncluttered search results model users loved, before Google stole its clothes and its thunder. It also introduced genuinely helpful online extras like advanced search operators, translations, and free ISP-agnostic webmail for those who didn’t want Hotmail. All this, before it was killed by mismanagement and becoming bloated.

One of the latest points of "what if" discussions surrounding AI advancements and their need to absorb every watt of computing power revolves around putting datacenters in space for power and cooling benefits, and itvsounds like a grand idea on its face.

But going by the thoughts and online calculator of Andrew McCalip from Varda Space Industries, his personal opinion is that the economics of such a proposition are a proverbial kick in the teeth, and should make any sensible observer ask "why?". A technical deep-dive by Andrew Côté from Hyperstition Incorporated essentially concludes the same.

On the face of it, the idea has merit — solar power is plentiful when you don't have to deal with that annoying day/night cycle, and space is pretty darn chilly, minus some "minor" engineering challenges. Even with extremely optimistic projections, plus the hope that reusable rockets will continue to evolve, and ignoring many associated costs, McCalip's math puts a SpaceGPT installation at over three times the cost of its earthbound equivalent: a baseline $51.1 billion for space, compared to $15.9 billion on terra firma.

His online calculator is extremely detailed, letting one configure 15 parameter sliders to try and get a feel for how good (or bad) the idea is. Said parameters range from adjusting the launch cost to Low Earth Orbit (LEO), satellite size, GPU failure rates, as well as comparable earthbound figures.

It should be noted that McCalip's math only accounts for the bare logistical differences of a launch versus a standard build, with a lot of assumptions about engineering challenges, and he remarks as much. McCalip goes as far as calling the notion "FOMO and aesthetic futurism", and that "people are using back-of-the-envelope math, doing a terrible job of it, and only confirming whatever conclusion they already want."

That's a shot across the bow at the (literal) pie-in-the-sky ideas of using LEO satellites to host datacenters. It's also easy to understand the frustration of those who wish that datacenter buildouts would move faster. For terrestrial data centers, tons of time is spent navigating legal and literal ground, worrying about water supplies, ecological and carbon output concerns, and so on and so forth.

Solar power is always available, so its effective price can approach near zero, helping with lowering the operational cost. There's no regulation regarding land (or barely any regulation at all), and no need to worry about passing transmission lines either. Cooling looks easy on the surface level, with space radiators being extremely efficient. Plus, with the advancements coming from SpaceX et al are reportedly hovering around $1000 per kg to launch into LEO.

Even though McCalip's analysis already puts the kibosh on the financial side, Andrew Côté's exceedingly deep dive into the engineering difficulties faced by such an enterprise delivers the killing blow. That's also an excellent read , as the list of things that can go wrong with setting up a datacenter in space can be reasonably summed up with "yes." Did you know, for example, that AI accelerators tend to up and die a lot when used at full tilt, and can't exactly be replaced in 5 minutes in space?

In these times where AI market valuations are climbing ever higher, it's arguably refreshing to see some down-to-earth thinking from those in know. Do check out McCalip's economic analysis and Côté's thoughts on the engineering challenges to be faced. Maybe one day the numbers will add up, but for now, it seems AI computing is best performed on good old terra firma.

Cloudflare CEO Matthew Prince said that his company has blocked over 416 billion AI bot requests since making it the default option in July of this year after it announced the Content Independence Day initiative. Prince said in an interview with Wired that this feature allows website owners to block AI crawlers by default, unless the AI company pays them to gain access to their content.

“The business model of the internet has always been to generate content that drive traffic and then sell either things, subscriptions, or ads, Prince told Wired. “What I think people don’t realize, though, is that AI is a platform shift. The business model of the internet is about to change dramatically. I don’t know what it’s going to change to, but it’s what I’m spending almost every waking hour thinking about.”

While Cloudflare blocks almost all AI crawlers, there’s one particular bot it cannot block without affecting its customers’ online presence — Google. The search giant combined its search and AI crawler into one, meaning users who opt out of Google’s AI crawler won't be indexed in Google search results. “You can’t opt out of one without opting out of both, which is a real challenge — it’s crazy,” Prince continued. “It shouldn’t be that you can use your monopoly position of yesterday in order to leverage and have a monopoly position in the market of tomorrow.”

Human-generated content is crucial for AI companies to train their models on, as research has proven that AI models turn to slop when trained on AI-generated data. AI summaries have been proven to reduce traffic on websites — especially impacting those that heavily rely on visibility and views for ad revenue — but licensing deals can help offset this, helping online publications remain a viable source of income for creators and publishers.

Cloudflare will also benefit from a varied internet that hosts content from real humans. Its CEO told Wired that the company is aiming for a future where creators and businesses grow on a level playing field, as there would be more websites that need protection, resulting in more potential clients for Cloudflare. This makes it one of the biggest content delivery networks in the world, owning 79.9% of the market as of 2022. However, this also makes the internet vulnerable, like when a single misconfigured file knocked out a huge chunk of the web in November.

This highlights the current problem we have with the global web infrastructure, which relies on just a few big companies — AWS, Azure, Cloudflare, CrowdStrike, and Google, to name a few — to serve the entire planet. While these institutions have made things easier and more streamlined for corporations that rely on the internet, it also means that just one of these services going down will cost billions in losses and severe disruption across the globe.

Nvidia released a statement saying that it’s “delighted by Google’s success,” but said that it continues to supply AI chips to the company. The AI giant made the X post in an apparent response to the news that Meta is in talks with Google to rent its Cloud Tensor Processing Units (TPUs) in 2026 and then to purchase them the following year. This news resulted in an increase in stock prices for Alphabet, Google’s parent company, and Meta, but also saw Nvidia take a 3% hit, especially as the market probably saw this as a break on the AI giant’s monopoly on the AI chip market.

"Nvidia is a generation ahead of the industry — it's the only platform that runs every AI model and does it everywhere computing is done," the company said on X via its Nvidia Newsroom account. "Nvidia offers greater performance, versatility, and fungibility than ASICs, which are designed for specific AI frameworks or functions." Google's Cloud TPUs are ASICs — application-specific integrated circuits — designed for AI workloads like training and inference. On the other hand, Nvidia's AI GPUs are general-purpose machines designed for parallel processing that have been repurposed for AI workloads.

Nvidia's Blackwell GPUs are indeed superior to almost anything else on the market, especially in terms of versatility. Although they're primarily advertised for AI applications, they can also be used for high-performance computing, data analytics, graphics rendering and visualization, and many more. On the other hand, Google's TPUs are designed for matrix multiplication — the arithmetic operation that primarily runs most modern AI. This means that you cannot easily repurpose these chips for other purposes. However, it also gives them an advantage in AI processing, even sometimes giving them an advantage in AI workloads, especially in terms of efficiency.

Meta isn't the first company to consider using Google TPUs for its AI operations. In fact, Anthropic has been using it since 2023, and the company has signed a deal to expand its contracted capacity to 1GW in 2026. But the biggest challenge for Google is the widespread adoption of its TPU chips, especially given Nvidia's huge market share. More than that, its CUDA platform is widely used in the industry, making it harder for developers to switch to Google's alternatives.

However, the size of the proposed deal between Meta and Google appears large enough to threaten Nvidia's dominance in the market. While it would still be a drop in the bucket compared to Nvidia's skyrocketing revenues, it would at least show potential clients that there's a viable alternative to Team Green's GPUs. Just this thought is enough to send Nvidia's stock price sliding, and it comes so soon after the company's record highs, leading CEO Jensen Huang to say the market does not appreciate the AI GPU manufacturer.

As the worldwide AI gold rush continues apace, the United States government continues to signal an urgent need for American leadership in this still-murky field. Yesterday, President Trump signed a new executive order outlining what the administration calls the Genesis Mission, an initiative that seeks to spur a nationwide effort to make the U.S. the world leader in AI technology and its practical application.

The White House's press release emphasizes the importance of its existing AI Action Plan, and compares the scale of its effort to the Manhattan Project, the initiative that produced the atomic bomb during World War II.

According to the new executive order, the plan is to be carried out by the Department of Energy (DoE), which controls key supercomputing resources in the USA among its other responsibilities. Secretary Chris Wright states: "The Genesis Mission will dramatically accelerate scientific discovery, strengthen national security, secure energy dominance, enhance workforce productivity, and multiply the return on taxpayer investment into research and development." There's no word on how the Genesis Mission is meant to be funded.

In its ideal form, the project is tasked with building an AI platform to collect federal scientific datasets to train AI models for the practical goal of solving at least 20 technological challenges. These will be collected among the areas of materials, manufacturing, biotechnology, nuclear fusion and fission energy, quantum information, semiconductors, and electronics.

Given the Genesis Mission is meant to be coordinated among several agencies and private partners, additional challenges can be proposed by those stakeholders. The challenges will be reviewed and adjusted yearly.

The first stage of the project will be to identify logical and physical resources, namely computing infrastructure, datasets, and models. The Genesis Mission could to make use of the supercomputers at DoE's national laboratories, but will also collaborate with agencies and partners "possessing advanced AI, data, or computing capabilities or scientific domain expertise."

That's a slightly curious statement, as most any capacity for datacenter hardware and AI chip manufacturing is de facto signed for for multiple years, meaning it's unclear whether the DoE wants to build out its own datacenters, rent existing capacity, or, most likely, both. An unnamed source at the White House told the New York Times that Nvidia, AMD, HP, and Dell have agreed to build facilities within national laboratories.

As for model training data, the project wants to ensure access to "appropriate datasets, including proprietary, federally curated, and open scientific datasets, in addition to synthetic data generated through DOE computing resources". Interestingly, the document makes repeated mentions of intellectual property protection and provenance tracking, presumably in the context of public-private partnerships, including "innovations arising from AI-directed experiments."

A program of this scale adds to the growing concerns about dwindling supplies and high prices for electricity, prompting the DoE Secretary to state the project should “make [the U.S.'] electricity grid more efficient and reverse price rises that have infuriated American citizens." The matching press release at energy.gov clearly states the Mission "will accelerate advanced nuclear, fusion, and grid modernization using AI to provide affordable, reliable, and secure energy for Americans."

There are no details on exact grid and generation upgrades, though that sentence seems to imply there could be a welcome nationwide effort to upgrade the nation's aging power infrastructure, and potentially build additional nuclear power plants and/or renewable energy farms.

The press release also makes repeated mentions of concerns with cybersecurity, an understandable notion given that how the main opponent is China, and the global supply chain of AI chips and datacenter hardware is, for now, manufactured or assembled in the region.

The Genesis Mission is an interesting proposition, especially in the face of the 15% export tax applied on Nvidia and AMD chips, and the fact that the U.S. government now holds 10% of Intel. On the one hand, it can be argued that there's no government AI effort that matches what companies like OpenAI, Google, and xAI have achieved—or at least, any made public. However, the project clearly outlines its intention to participate with industry partners, likely giving them a boost up the AI road, even if just with funding.

Given that even Nvidia's leather-clad man Jensen Huang believes that China is about to be ahead in the AI race, perhaps it's high time the U.S. government got its sneakers out.