China's racing to build its AI ecosystem as U.S. tech curbs bite. Here's how its supply chain stacks up
With the U.S. restricting China from buying advanced semiconductors used in artificial intelligence development, Beijing is placing hopes on domestic alternatives such as Huawei.
The task has been made more challenging by the fact that U.S. curbs not only inhibit China's access to the world's most advanced chips, but also restrict availing technology vital for creating an AI chip ecosystem.
Those constraints span the entire semiconductor value chain, ranging from design and manufacturing equipment used to produce AI chips to supporting elements such as memory chips.
Beijing has mobilized tens of billions of dollars to try to fill those gaps, but while it has been able to "brute force" its way into some breakthroughs, it still has a long way to go, according to experts.
"U.S. export controls on advanced Nvidia AI chips have incentivized China's industry to develop alternatives, while also making it more difficult for domestic firms to do so," said Paul Triolo, partner and senior vice president for China at advisory firm DGA-Albright Stonebridge Group.
Here's how China stacks up against the rest of the world in four key segments needed to build AI chips.
Nvidia is regarded as the world's leading AI chip company, but it's important to understand that it doesn't actually manufacture the physical chips that are used for AI training and computing.
Rather, the company designs AI chips, or more precisely, graphics processing units. Orders of the company's patented GPU designs are then sent to chip foundries — manufacturers that specialize in the mass production of other companies' semiconductor products.
While American competitors such as AMD and Broadcom offer varying alternatives, GPU designs from Nvidia are widely recognized as the industry standard. The demand for Nvidia chips is so strong that Chinese customers have continued to buy any of the company's chips they can get their hands on.
But Nvidia is grappling with Washington's tightening restrictions. The company revealed in April that additional curbs had prevented it from selling its H20 processor to Chinese clients.
Nvidia's H20 was a less sophisticated version of its H100 processor, designed specifically to skirt previous export controls. Nevertheless, experts say, it was still more advanced than anything available domestically. But China hopes to change that.
In response to restrictions, more Chinese semiconductor players have been entering the AI processor arena. They've included a wide array of upstarts, such as Enflame Technology and Biren Technology, seeking to soak up billions of dollars in GPU demand left by Nvidia.
But no Chinese firm appears closer to providing a true alternative to Nvidia than Huawei's chip design arm, HiSilicon.
Huawei's most advanced GPU in mass production is its Ascend 910B. The next-generation Ascend 910C was reportedly expected to begin mass shipments as early as May, though no updates have emerged.
Dylan Patel, founder, CEO and chief analyst at SemiAnalysis, told CNBC that while the Ascend chips remain behind Nvidia, they show that Huawei has been making significant progress.
"Compared to Nvidia's export-restricted chips, the performance gap between Huawei and the H20 is less than a full generation. Huawei is not far behind the products Nvidia is permitted to sell into China," Patel said.
He added that the 910B was two years behind Nvidia as of last year, while the Ascend 910C is only a year behind.
But while that suggests China's GPU design capabilities have made great strides, design is just one aspect that stands in the way of creating a competitive AI chip ecosystem.
To manufacture its GPUs, Nvidia relies on TSMC, the world's largest contract chip foundry, which produces most of the world's advanced chips.
TSMC complies with U.S. chip controls and is also barred from taking any chip orders from companies on the U.S. trade blacklist. Huawei was placed on the list in 2019.
That has led to Chinese chip designers like Huawei to enlist local chip foundries, the largest of which is SMIC.
SMIC is far behind TSMC — it's officially known to be able to produce 7-nanometer chips, requiring less advance tech than TSMC's 3-nanometer production. Smaller nanometer sizes lead to greater chip processing power and efficiency.
There are signs that SMIC has made progress. The company is suspected to have been behind a 5-nanometer 5G chip for Huawei's Mate 60 Pro, which had rocked confidence in U.S. chip controls in 2023. The company, however, has a long way to go before it can mass-produce advanced GPUs in a cost-efficient manner.
According to independent chip and technology analyst Ray Wang, SMIC's known operation capacity is dwarfed by TSMC's.
"Huawei is a very good chip design company, but they are still without good domestic chipmakers," Wang said, noting that Huawei is reportedly working on its own fabrication capabilities.
But the lack of key manufacturing equipment stands in the way of both companies.
SMIC's ability to fulfill Huawei's GPU requirements is limited by the familiar problem of export controls, but in this case, from the Netherlands.
While Netherlands may not have any prominent semiconductor designers or manufacturers, it's home to ASML, the world's leading supplier of advanced chipmaking equipment — machines that use light or electron beams to transfer complex patterns onto silicon wafers, forming the basis of microchips.
In accordance with U.S. export controls, the country has agreed to block the sale of ASML's most advanced ultraviolet (EUV) lithography machines. The tools are critical to making advanced GPUs at scale and cost-effectively.
EUV is the most significant barrier for Chinese advanced chip production, according to Jeff Koch, an analyst at SemiAnalysis. "They have most of the other tooling available, but lithography is limiting their ability to scale towards 3nm and below process nodes," he told CNBC.
SMIC has found methods to work around lithography restrictions using ASML's less advanced deep ultraviolet lithography systems, which have seen comparatively fewer restrictions.
Through this "brute forcing," producing chips at 7 nm is doable, but the yields are not good, and the strategy is likely reaching its limit, Koch said, adding that "at current yields it appears SMIC cannot produce enough domestic accelerators to meet demand."
SiCarrier Technologies, a Chinese company working on lithography technology, has reportedly been linked to Huawei.
But imitating existing lithography tools could take years, if not decades, to achieve, Koch said. Instead, China is likely to pursue other technologies and different lithography techniques to push innovation rather than imitation, he added.
While GPUs are often identified as the most critical components in AI computing, they're far from the only ones. In order to operate AI training and computing, GPUs must work alongside memory chips, which are able to store data within a broader "chipset."
In AI applications, a specific type of memory known as HBM has become the industry standard. South Korea's SK Hynix has taken the industry lead in HBM. Other companies in the field include Samsung and U.S.-based Micron.
"High bandwidth memory at this stage of AI progression has become essential for training and running AI models," said analyst Wang.
As with the Netherlands, South Korea is cooperating with U.S.-led chip restrictions and began complying with fresh curbs on the sale of certain HBM memory chips to China in December.
In response, Chinese memory chip maker ChangXin Memory Technologies, or CXMT, in partnership with chip-packaging and testing company Tongfu Microelectronics, is in the early stages of producing HBM, according to a report by Reuters.
According to Wang, CXMT is expected to be three to four years behind global leaders in HBM development, though it faces major roadblocks, including export controls on chipmaking equipment.
SemiAnalysis estimated in April that CXMT remained a year away from ramping any reasonable volume.
Chinese foundry Wuhan Xinxin Semiconductor Manufacturing is reportedly building a factory to produce HBM wafers. A report from SCMP said that Huawei Technologies had partnered with the firm in producing HBM chips, although the companies did not confirm the partnership.
Huawei has leaned on HBM stockpiles from suppliers like Samsung for use in their Ascend 910C AI processor, SemiAnalysis said in an April report, noting that while the chip was designed domestically, it still relies on foreign products obtained prior to or despite restrictions.
"Whether it be HBM from Samsung, wafers from TSMC, or equipment from America, Netherlands, and Japan, there is a big reliance on foreign industry," SemiAnalysis said.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
14 minutes ago
- Yahoo
What Are the Top 5 Artificial Intelligence (AI) Stocks to Buy Right Now?
Nvidia and Broadcom are two of the best plays on the artificial intelligence (AI) infrastructure build-out. TSMC has been a vital part of the semiconductor value chain and is set to benefit from increasing AI chip production. Palantir Technologies and GitLab are two strong AI software plays. 10 stocks we like better than Nvidia › Artificial intelligence (AI) has the potential to be the most important technological advancement in history, and it still appears to be in its early innings. As such, the space is still one of the most promising places to invest. When delving down to specific stocks to focus attention on, five of the best AI-related offerings to buy right now are Nvidia (NASDAQ: NVDA), Broadcom (NASDAQ: AVGO), Taiwan Semiconductor Manufacturing (NYSE: TSM), Palantir Technologies (NASDAQ: PLTR), and GitLab (NASDAQ: GTLB). Each company has found a unique niche in the AI sector to exploit, and each has big opportunities ahead. Let's look at why these are among the best AI stocks to invest in right now for the long term. Nvidia's graphics processing units (GPUs) have become the backbone of AI infrastructure. These powerful chips are used to help power AI workloads, and demand has been soaring. The company's wide moat, however, comes from its CUDA software platform. Nvidia launched CUDA two decades ago to allow developers to program its GPUs more easily. It pushed CUDA into universities and research institutions early on, which helped CUDA become the de facto software program for AI developers. In Q1, the company had an over 90% market share in the GPU space. As a result, where AI infrastructure spending goes, Nvidia is sure to follow. While a slowdown in data center spending would be a risk, right now, demand for AI chips is only getting bigger. This is being driven by cloud computing companies pouring money into AI infrastructure to keep up with demand; large tech companies and AI start-ups spending big to create new foundational AI models; and even countries making large investments to not fall behind in the AI race. As AI infrastructure spending continues to ramp up, Nvidia remains a clear winner. Another company taking advantage of the AI infrastructure build-out is Broadcom. Rather than designing GPUs like Nvidia, it's been focused on networking components and helping customers design custom AI chips. It also added a software component when it acquired VMWare. Thus far, its strategy is paying off. Its Ethernet switches and other networking components help efficiently move data within huge AI clusters, making them an essential part of data center infrastructure. Last quarter, its AI networking revenue soared 70% and accounted for 40% of Broadcom's total AI revenue. However, Broadcom's biggest long-term opportunity comes from helping customers design custom AI chips, which can offer better performance and lower power consumption than off-the-shelf GPUs. Demand is starting to pick up. Broadcom says its top three custom chip customers are on track to deploy 1 million AI chip clusters each by 2027, representing a total opportunity of between $60 billion to $90 billion. On the software side, Broadcom also benefits from transitioning VMWare customers from perpetual licenses to a subscription models and upgrading them to its VMware Cloud Foundation (VCF) platform. VCF helps customers build hybrid and multi-cloud environments so they can manage workloads across public clouds and their own on-premises data centers. At the end of last quarter, 87% of its top 10,000 customers had adopted VCF. While an AI infrastructure spending slowdown is a risk, given its networking leadership, custom AI chip opportunity, and growing software revenue, Broadcom is well-positioned moving forward. While semiconductor companies garner most of the attention from investors, Taiwan Semiconductor Manufacturing is the company that actually manufactures most of these AI chips. It's the clear leader in advanced semiconductor manufacturing and a key partner to top customers like Nvidia, Apple, and Broadcom. Meanwhile, AI is driving its business. High-performance computing now makes up 59% of its revenue, up from 46% a year ago. Most of that comes from advanced nodes. Nodes refer to the manufacturing process used to make chips. The smaller the number (measured in nanometers), the more transistors you can pack onto a chip. This improves a chip's performance and power efficiency, and TSMC is the best in the world at making these at scale. And with rivals struggling to make advanced chips, this has given TSMC strong pricing power, as well. The biggest risk to TSMC is a slowdown in AI infrastructure spending, which would hit both revenue and fab utilization. However, the company is working closely with its largest customers to build out capacity in lockstep with their demand. With advanced-node capacity tight, TSMC is well-positioned to continue to be an AI infrastructure build-out winner. Palantir Technologies has emerged as a key player in the AI space. Instead of putting resources into developing AI models, the company focuses on the applications and workflow layers of AI to essentially develop an AI operating system. It does this by gathering data from a wide array of sources and organizing it into an ontology that links the data to its real-world counterparts. As a result, the company's AI Platform (AIP) can help organizations solve complex problems. This includes everything from monitoring sepsis in hospitals to streamlining underwriting processes in insurance. The sheer number of use cases across various industries that AIP can handle is just an enormous opportunity for Palantir moving forward. The company saw its revenue growth consistently accelerate over the past two years, including a 39% increase last quarter. The stock is not without risks, as it carries a high valuation and is exposed to government budget cuts, since the federal government is still its largest client. However, Palantir is unique in the AI space and has one of the biggest opportunities in front of it. GitLab is a leader in the DevSecOps space, offering a platform that helps developers build software securely. It's been an AI winner, as customers expand seats and upgrade to its higher-tier platforms as AI increases the customers' software development. GitLab is also helping customers become more productive with products like GitLab Duo, which uses AI to provide code suggestions and automation to streamline development. The company consistently delivered strong revenue growth of between 25% to 40% over the past two years and boasts impressive gross margins and solid free cash flow. It's also done a great job growing within its existing customer base, as evidenced by its 122% dollar-based net retention over the past 12 months. Most of this is coming from seat expansions, followed by strong upgrades to higher-tiered offerings. While there has been concern that AI will replace coders, right now the opposite has been true, as GitLab's technology is being used to make software developers more productive, not replace them. This is driving strong growth and makes the company look like a solid long-term winner. Before you buy stock in Nvidia, consider this: The Motley Fool Stock Advisor analyst team just identified what they believe are the for investors to buy now… and Nvidia wasn't one of them. The 10 stocks that made the cut could produce monster returns in the coming years. Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you'd have $659,171!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you'd have $891,722!* Now, it's worth noting Stock Advisor's total average return is 995% — a market-crushing outperformance compared to 172% for the S&P 500. Don't miss out on the latest top 10 list, available when you join . See the 10 stocks » *Stock Advisor returns as of June 9, 2025 Geoffrey Seiler has positions in GitLab. The Motley Fool has positions in and recommends Apple, GitLab, Nvidia, Palantir Technologies, and Taiwan Semiconductor Manufacturing. The Motley Fool recommends Broadcom. The Motley Fool has a disclosure policy. What Are the Top 5 Artificial Intelligence (AI) Stocks to Buy Right Now? was originally published by The Motley Fool Sign in to access your portfolio
Newsweek
an hour ago
- Newsweek
How Animal Testing in US Could Be Transformed Under Trump
Based on facts, either observed and verified firsthand by the reporter, or reported and verified from knowledgeable sources. Newsweek AI is in beta. Translations may contain inaccuracies—please refer to the original content. Millions of animals each year are killed in U.S. laboratories as part of medical training and chemical, food, drug and cosmetic testing, according to the non-profit animal rights organization People for the Ethical Treatment of Animals (PETA). For many animals held captive for research, including a huge range of species from dogs, cats and hamsters to elephants, dolphins and many other species, pain is "not minimized," U.S. Department of Agriculture data shows. The issue of animal testing is something most Americans agree on: it needs to change and gradually be stopped. A Morning Consult poll conducted at the end of last year found that 80 percent of the 2,205 participants either agreed or strongly agreed with the statement: "The US government should commit to a plan to phase out experiments on animals." Since President Donald Trump began his second term, his administration has been making moves to transform and reduce animal testing in country, although the question remains as to whether it will be enough to spare many more animals from pain and suffering this year. Animal Testing In US Could Be Transformed Animal Testing In US Could Be Transformed Photo-illustration by Newsweek/Getty/Canva What Is The Trump Administration Doing About It? There have been various steps taken in different federal agencies to tackle the issue of animal testing since Trump was sworn in on January 20. In April, the Food and Drug Administration (FDA) announced it was "taking a groundbreaking step to advance public health by replacing animal testing in the development of monoclonal antibody therapies and other drugs with more effective, human-relevant methods." The FDA said that its animal testing requirement will be "reduced, refined, or potentially replaced" with a range of approaches, including artificial intelligence-based models, known as New Approach Methodologies or NAMs data. A Department of Health and Human Services (HHS) official told Newsweek: "The agency is paving the way for faster, safer, and more cost-effective treatments for American patients. "As we restore the agency's commitment to gold-standard science and integrity, this shift will help accelerate cures, lower drug prices, and reaffirm U.S. leadership in ethical, modern science." The National Institutes of Health (NIH) announced it was "adopting a new initiative to expand innovative, human-based science while reducing animal use in research," in alignment with the FDA's initiative. The agency said that while "traditional animal models continue to be vital to advancing scientific knowledge," new and emerging technologies could act as alternative methods, either alone or in combination with animal models. The NIH Office of Extramural Research told Newsweek it was "committed to transparently assessing where animal use can be reduced or eliminated by transitioning to [new approach methodologies (NAMs)]." "Areas where research using animals is currently necessary represent high-priority opportunities for investment in NAMs," the agency added. It added that it will "further its efforts to coordinate agency-wide efforts to develop, validate, and scale the use of NAMs across the agency's biomedical research portfolio and facilitate interagency coordination and regulatory translation for public health protection." During Trump's first term, the Environmental Protection Agency (EPA) signed a directive to "prioritize efforts to reduce animal testing and committed to reducing testing on mammals by 30 percent by 2025 and to eliminate it completely by 2035," an EPA spokesperson told Newsweek. Although, the spokesperson added: "the Biden Administration halted progress on these efforts by delaying compliance deadlines." As a member of the House, Lee Zeldin, the EPA's current administrator, co-sponsored various bills during Trump's first term regarding animal cruelty, covering issues such as phasing out animal-based testing for cosmetic products; ending taxpayer funding for painful experiments on dogs at the Department of Veteran Affairs; empowering federal law enforcement to prosecute animal abuse cases that cross state lines; and others, the spokesperson said. What The Experts Think Needs To Be Done The Trump administration's efforts to tackle the issue of animal testing appear to be a step in the right direction, according to experts who spoke with Newsweek. "I was pleasantly surprised and quite frankly a bit shocked to read the simultaneous announcements by the NIH and the FDA regarding a new emphasis on the use of alternatives to animals," Jeffrey Morgan, a professor of pathology and laboratory medicine at Brown University in Rhode Island, told Newsweek. Morgan, who is also the director of the Center for Alternatives to Animals in Testing at Brown University, said that both agencies are moving together in the same direction on the issue "sends a unified and very powerful message to the research and biotech communities." He added that the announcements showed "a major acknowledgement of the limitations of the use of animals in research and testing." "What is especially exciting is that the NIH announcement will encourage the entry of new investigators into the field, further accelerating innovation in alternatives with exciting impacts for both discovery and applied research across all diseases," he said. He added that the FDA announcement and its emphasis on a new regulatory science that embraces data from alternatives was "equally exciting." "The demands of this new regulatory science will likewise accelerate innovation because it will establish the much-needed regulatory framework for the rigorous evaluation of data from alternatives," he said. While the administration's initiatives to shift research away from animal testing is heading in the right direction, its policies are "overdue," Dr. Thomas Hartung, a professor in the department of environmental health and engineering at Johns Hopkins Bloomberg School of Public Health, Maryland, told Newsweek. "The animal tests for safety were introduced more than 50 years ago. There is no other area of science where we do not adapt to scientific progress," he said. Hartung added that animal "testing takes too long and is too expensive to really provide the safety consumers want." He said that running animal tests for new chemicals can cost millions and take years in some cases. "Nobody can wait that long, even if they can afford the testing costs," he said. Hartung also believes the shifts in the industry to reduce animal testing have been "coming for a while," as over the last two decades, America's opposition to animal use in medical research has been increasing. "The alignment of FDA and NIH really makes the difference now, which I think is evidence of a strong relationship of their leaderships," he said. Yet in order to make a real difference, Hartung said clear deadlines are key to show that "this is not just lip service." He also said that he thought "the transformative nature of artificial intelligence in this field is not fully acknowledged." "We also need an objective framework for change to better science, such as the evidence-based toxicology approach," he said.
Yahoo
2 hours ago
- Yahoo
Huawei Patents Solid-State Battery That Would Let You Road Trip From Miami To Maine On A Single Charge
Toyota is currently working on solid-state batteries it says will allow it to build 900-mile EVs, but while many would call that overkill, dads across the country know it isn't enough. We need more range if EVs are ever going to be good enough for a proper road trip, and if Huawei can deliver on its latest solid-state battery patent, we may almost be there. That's because it offers more than 1,800 miles of range on a single charge, CarNewsChina reports. For context, that's enough range to drive from Miami, Florida, to Bangor, Maine, without stopping. That said, China's range estimates are even more generous than Europe's, which means you may only get 1,200 or 1,300 miles out of a battery like that in the United States. Pitiful, I know. At that point, there's basically no reason to even bring your external catheter on a road trip. Not that the U.S. would allow Chinese battery tech to be sold here in the first place or invest in catching up. Our EV tech will stay generations behind China's, and we will like it. From the sound of it, the real value in this patent is the battery's energy density: The patent outlines a solid-state battery architecture with energy densities between 400 and 500 Wh/kg, potentially two to three times that of conventional lithium-ion cells. The filing also details a novel approach to improving electrochemical stability: doping sulfide electrolytes with nitrogen to address side reactions at the lithium interface, a long-standing obstacle to the commercialisation of sulfide-based batteries. Huawei's design aims to boost safety and cycle life by mitigating degradation at this critical junction. Read more: These Supercars Lose Value So Quickly, They're Almost A Steal Allegedly, given a big enough charger, Huawei says it could charge the battery in about five minutes, too. That would require an absolutely absurd amount of power, but it isn't like companies in China aren't working on that, too. BYD has already shown off a Megawatt charger that's capable of adding 250 miles of range in about five minutes. Early versions would also likely cost a small fortune, though, and it's also hard to see how practical Megawatt charging would be in the remote areas where people might actually use it. But while BYD does actually plan to build at least some of those Megawatt chargers, don't be surprised if Huawei's 1,800-mile solid-state battery ends up being more of a tech demonstration than anything else. Instead, it seems much more likely that the batteries that will ultimately make their way into production cars will be smaller and therefore lighter. A battery that's a third the size of the 1,800-mile battery would still offer more than enough range for a road trip while also drastically reducing weight. Considering how heavy EV batteries are — the battery in the Hummer EV weighs nearly 3,000 pounds — if there's one thing EVs need, it's less weight. So while Huawei's 1,800-mile battery is still theoretical for now, it's still cool to see battery tech advancing to the point where, in a few years, we may see long-range EVs that weigh about the same as gas-powered cars. At least in other countries outside the U.S., since Republicans have decided they'd rather just turn us into a technological backwater. Want more like this? Join the Jalopnik newsletter to get the latest auto news sent straight to your inbox... Read the original article on Jalopnik.
