Chinese memory and storage firm expected to post more than 60,000% jump in profits due to exploding demand — Lexar owner Longsys forecasts nearly $1.5 billion profit for 1H26 compared to $2.1 million last year
The memory and chip shortage is driving record profits for downstream memory suppliers.
Shenzhen Longsys Electronics, the Chinese parent company of Lexar, announced that it’s expecting a net profit of $1.36 to $1.62 billion (or 9.2 to 11 billion yuan) for the first half of 2026, smashing its year-ago profit of just $2.2 million. This represents an incredible jump of 61,818% to 73,636%, and it comes on revenue forecasts of $3.24 to $3.68 billion (22 to 25 billion yuan) — more than double the $1.5 billion it achieved last year, according to the South China Morning Post (SCMP).
This massive growth is attributed to the increased demand for—you guessed it— memory and storage chips due to the global AI infrastructure buildout, and all that demand is competing for limited memory wafer capacity. Longsys says that it has signed long-term agreements and memoranda of understanding with global memory wafer suppliers to ensure supply stability.
The company did not specifically say which memory and storage chip suppliers it has contracted with, but many Chinese memory brands have been ditching the three mainstream suppliers — Micron, Samsung, and SK hynix — for CXMT and YMTC silicon. U.S. manufacturers like Corsair, Dell, and HP, have started considering chips from these suppliers, despite being labeled as Chinese military companies by the Pentagon. Even Apple, which used to have massive sway on its suppliers, has started lobbying Washington for access to CXMT chips as Samsung and SK hynix say that the AI-driven shortages could last until 2027 or even longer.
Longsys’ unprecedented growth has resulted in a 12.5% jump on its stock price in the Shenzhen Stock Exchange over the weekend, which has more than doubled from its lowest point just three months ago. Aside from its record earnings, the company also received the go-ahead from Chinese regulators to raise up to $544 million (3.7 billion yuan) through a private share placement. This would allow it to directly offer shares to select investors and fund research and development on high-end memory products, including AI-focused storage solutions, storage and memory controllers, among others.
The memory and chip shortage is hurting PC makers and consumers, with the PC market expected to shrink by 14% this year. As the market becomes desperate for alternative sources to mainstream manufacturers, Chinese firms CXMT and YMTC, as well as other downstream suppliers like Longsys and Biwin, are taking the opportunity to challenge established Western brands.
Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.
Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.

Jowi Morales is a tech enthusiast with years of experience working in the industry. He’s been writing with several tech publications since 2021, where he’s been interested in tech hardware and consumer electronics.
-
bit_user So, if one year they barely broke even, turning a $1 profit, and the next year they made $1M of profit, their profits would've grown 1 Million Times!!! Even though maybe their revenues increased only slightly (or maybe their costs decreased and their revenues just decreased a bit less).Reply
LOL. -
bit_user Reply
True. But, it's a cyclical business. You'd better take the profits in the fat times, because the lean times can be brutal. That's why so many other DRAM makers have gone under, gotten bought out, or gotten out of the business.TechieTwo said:Proof of price gouging - because they can. :( -
SkyBill40 And next to none of this financial boon will be shared with those workers actually doing the labor.Reply -
Staying RightBehindYou Reply
I don't think it will ever be back to normal.ivan_vy said:Well, my DDR4 memory will have to last another five years, at the very least. -
usertests Reply
Even in the absence of the AI boom, there has always been pressure to deliver denser memory, which lowers manufacturing cost per bit. Around 2022-2024, 24-32 Gb memory dies were introduced. 16 Gb had been around since 2018.Staying RightBehindYou said:I don't think it will ever be back to normal.
Samsung and others are working on transitioning to vertical channel transistors (VCTs) within the next few years, which could allow for 48-64 Gb dies. During the 2030s, there will be a transition to monolithic 3D DRAM, for dies that might start out in the 128-192 Gb range, but could eventually scale to >1 terabit. NEO Semiconductor is trying to beat the usual suspects to the punch with 3D X-DRAM.
So there will certainly come a time when DRAM will fall back to, and even below, whatever you consider "normal". I assume $2-3/GB? And it doesn't require the memory manufacturers to be selling the chips at a loss, if they increased density during that time.
There is an assumption that AI will gobble up memory forever. There have been many memory boom and bust cycles in the past. I think it's safe to assume that the pattern will continue, with an upcoming bust. And you can find plenty of commentary about the AI bubble, companies recently dialing back on their token spending, etc.
Maybe it will end up taking over a decade, but it will happen. -
ivan_vy Replyusertests said:Even in the absence of the AI boom, there has always been pressure to deliver denser memory, which lowers manufacturing cost per bit. Around 2022-2024, 24-32 Gb memory dies were introduced. 16 Gb had been around since 2018.
Samsung and others are working on transitioning to vertical channel transistors (VCTs) within the next few years, which could allow for 48-64 Gb dies. During the 2030s, there will be a transition to monolithic 3D DRAM, for dies that might start out in the 128-192 Gb range, but could eventually scale to >1 terabit. NEO Semiconductor is trying to beat the usual suspects to the punch with 3D X-DRAM.
So there will certainly come a time when DRAM will fall back to, and even below, whatever you consider "normal". I assume $2-3/GB? And it doesn't require the memory manufacturers to be selling the chips at a loss, if they increased density during that time.
There is an assumption that AI will gobble up memory forever. There have been many memory boom and bust cycles in the past. I think it's safe to assume that the pattern will continue, with an upcoming bust. And you can find plenty of commentary about the AI bubble, companies recently dialing back on their token spending, etc.
Maybe it will end up taking over a decade, but it will happen.
AI will eventually hit a usability ceiling—reaching a point where it is "good enough"—and find a balance between investment and return. I don't think it will keep consuming hardware at this same rate forever; it needs to stabilize, just like any other server-based services. We are currently in the "selling shovels" phase, happened with GPUs and cryptos, with storage (Chia crypto) and the covid hitting car manufactures with low IC production ; but the issue is that we need those shovels in our phones, cars, and personal computers as well.