Connect with us

Published

on

Chip engineer handling a wafer.

Sinology | Moment | Getty Images

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. 

AI chip design

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.

AI chip fabrication

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.

Advanced Chip equipment  

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.

AI memory components

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.

Continue Reading

Technology

Winklevoss-founded Gemini reportedly prices IPO at $28 per share, valuing the crypto exchange at $3.3 billion

Published

on

By

Winklevoss-founded Gemini reportedly prices IPO at  per share, valuing the crypto exchange at .3 billion

Tyler Winklevoss and Cameron Winklevoss (L-R), creators of crypto exchange Gemini Trust Co., on stage at the Bitcoin 2021 Convention, a cryptocurrency conference held at the Mana Convention Center in Wynwood in Miami, Florida, on June 4, 2021.

Joe Raedle | Getty Images

Gemini Space Station, the crypto company founded by Cameron and Tyler Winklevoss, priced its initial public offering at $28 per share late Thursday, according to Bloomberg.

A person familiar with the offering told the news service that the company priced the offering above its expected range of $24 to $26, which would value the company at $3.3 billion.

Since Gemini capped the value of the offering at $425 million, 15.2 million shares were sold, according to the report. That was a measure of high demand for the crypto company, which had initially marketed 16.67 million shares. Earlier this week, it increased its proposed price range from between $17 and $19 apiece.

A Gemini spokesperson could not confirm the report.

The company and the selling stockholders granted its underwriters — led by and Goldman Sachs, Citigroup and Morgan Stanley — a 30-day option to sell an additional 452,807 and 380,526 shares, respectively, per the registration form. Gemini stock will trade on the Nasdaq under ticker symbol “GEMI.”

Up to 30% of the shares offered will be reserved for retail investors through Robinhood, SoFi, Hong Kong-based Futu Securities, Singapore’s Moomoo Financial, Webull and other platforms.

Gemini, which primarily operates as a cryptocurrency exchange, was founded by the Winklevoss brothers in 2014 and holds more than $21 billion of assets on its platform as of the end of July.

Initial trading will give the market a sense of how long it can keep the crypto IPO party going. Circle Internet and Bullish had successful listings, but there has been a recent consolidation in the prices of blue chip cryptocurrencies like bitcoin and ether. Also, in contrast to those companies’ profitability, Gemini has reported widening losses, especially in 2025. Per its registration with the Securities and Exchange Commission, Gemini posted a net loss of $159 million in 2024, and in the first half of this year, it lost $283 million.

This week, however, Gemini received a big vote of institutional confidence when Nasdaq said it’s making a strategic investment of $50 million in the crypto company. Nasdaq is seeking to offer its clients access to Gemini’s custodial services, and gain a distribution partner for its trade management system known as Calypso.

Gemini also offers a crypto-backed credit card, and last month, launched another card in partnership with Ripple. The latter garnered more than 30,000 credit card sign-ups in August, a new monthly high that was more than twice the number of credit card sign-ups in the prior month, according to the S-1 filing.

Don’t miss these cryptocurrency insights from CNBC Pro:

(Learn the best 2026 strategies from inside the NYSE with Josh Brown and others at CNBC PRO Live. Tickets and info here.)

Continue Reading

Technology

OpenAI says nonprofit parent will own equity stake in company of over $100 billion

Published

on

By

OpenAI says nonprofit parent will own equity stake in company of over 0 billion

Microsoft Chairman and Chief Executive Officer Satya Nadella (L), speaks with OpenAI Chief Executive Officer Sam Altman, who joined by video during the Microsoft Build 2025, conference in Seattle, Washington on May 19, 2025.

Jason Redmond | AFP | Getty Images

OpenAI on Thursday said its nonprofit parent will continue to have oversight over the company and will own an equity stake of more than $100 billion.

The artificial intelligence startup, recently valued at $500 billion, said this structure will make the nonprofit “one of the most well-resourced philanthropic organizations in the world,” and will allow the company to continue to raise capital.

OpenAI also announced it has signed a non-binding memorandum of understanding with Microsoft, which outlines the next phase of their partnership. Microsoft has invested over $13 billion in OpenAI, backing the company as early as 2019, three years before the launch of of the chatbot ChatGPT.

“We are actively working to finalize contractual terms in a definitive agreement,” OpenAI said in a joint statement with Microsoft, which is also the company’s key cloud partner. “Together, we remain focused on delivering the best AI tools for everyone, grounded in our shared commitment to safety.”

In May, OpenAI bowed to pressure from civic leaders and ex-employees, announcing that its nonprofit would retain control even as the company was restructuring into a public benefit corporation. OpenAI was founded as a nonprofit research lab in 2015, but has in recent years become one of the fastest-growing commercial entities on the planet.

OpenAI said Thursday it is working closely with the California and Delaware Attorneys General to establish its structure.

“OpenAI started as a nonprofit, remains one today, and will continue to be one – with the nonprofit holding the authority that guides our future,” the company’s Chairman Bret Taylor said in a statement Thursday.

The startup has been engulfed in a heated legal battle with Elon Musk, one of its co-founders. Musk has been trying to keep OpenAI from converting into a for-profit company as he competes in the generative AI market with his own startup, xAI.

OpenAI said its nonprofit is also opening applications for the first phase of a $50 million grant initiative that is aimed to support other nonprofit and community organizations across AI literacy, economic opportunity and community innovation.

WATCH: AI’s trillion-dollar money loop

AI's trillion dollar money loop

Continue Reading

Technology

‘We will do better.’ Microsoft CEO Nadella admits company has to rebuild trust with employees

Published

on

By

'We will do better.' Microsoft CEO Nadella admits company has to rebuild trust with employees

Microsoft CEO Satya Nadella departs following a meeting of the White House Task Force on AI Education in the East Room of the White House in Washington on Sept. 4, 2025.

Eric Lee | Bloomberg | Getty Images

Microsoft CEO Satya Nadella told employees in a meeting on Thursday that the company has work to do to smooth relations with employees after announcing several rounds of layoffs and a mandated partial return to in-person work.

In the meeting that was held online, an employee asked executives to speak about a perceived lack of empathy in the company’s culture as of late and steps Microsoft is taking to rebuild trust with its workforce.

“I deeply appreciate that, the question and the sentiment behind it,” Nadella said, in audio that was obtained by CNBC. “I take it as feedback for me and everyone in the leadership team, because at the end of the day, I think we can do better, and we will do better.”

Nadella’s comments come after Microsoft slashed 9,000 jobs in July, following smaller reductions in the months prior. On Tuesday, Microsoft said workers living near its headquarters in Redmond, Washington, must come into the office three days a week, starting in February, with a broader rollout to follow.

Amy Coleman, Microsoft’s human resources chief, said at Thursday’s meeting that reception to the return-to-office announcement has been mixed, with some workers feeling like they’re losing autonomy. But she said that employees in and around Seattle already come in, on average, 2.4 times each week.

Like most of the tech industry, Microsoft went fully remote during the pandemic, and made particular use of its internal Teams video and chat offerings, which gained rapid adoption during that period. Microsoft has been slower than many of its peers to put a mandate in place for coming back to the office. Amazon, one of Microsoft’s top rivals, called employees back to offices five days a week in January.

While Nadella and the executive team are taking criticism from some staffers, Wall Street is applauding the company’s growth and execution. The stock is up almost 20% this year, outperforming the broader market, pushing Microsoft’s market cap to $3.7 trillion, which trails only Nvidia among the world’s most-valuable companies.

In July, Microsoft reported a 24% increase in net income to $27 billion. The company’s gross margin was under 69%, compared with 71% in late 2023. It’s rapidly building and renting data center infrastructure to meet artificial intelligence demand.

AI infrastructure build-up is a long-term story as adoption is only consumer based now

Nadella said at the meeting that with remote work, new employees and those who are early in their careers don’t always feel a sense of apprenticeship or mentorship.

“Management is just mostly all remote, but the interns are all, you know, in one location,” he said. “And so those are things that just will break a social contract.”

Microsoft didn’t immediately provide a comment.

Even with Microsoft’s rapid expansion, Nadella said the company is feeling the pressure. It’s a common theme in the software industry, as concerns proliferate about the impact of AI and its potential to automate work.

“We have some very, very hard work ahead of us, and that hard process of renewal is essentially what we have to do,” Nadella said. “You have to be hardcore in terms of an intellectual honesty about what really needs to happen.”

Microsoft’s Azure cloud business grew 39% in the latest quarter, but revenue in the Windows and devices business increased by just 2.5%.

“Some of the biggest businesses we built may not be as relevant going forward,” Nadella said. “Some of the margin that we love today may not be there tomorrow, and that means you have to be way ahead of all of those going away, right?”

Microsoft, which celebrated its 50th anniversary in April, will retain its core values as it confronts market realities, Nadella said.

“Capital markets have one simple truth,” he said. “There is no permission for any company to exist forever.”

That wasn’t the only contentious topic at the meeting.

Employees are awaiting details from a third-party investigation after The Guardian said in August that Israel’s military used Microsoft’s Azure cloud infrastructure to store Palestinians’ phone calls as part of Israel’s invasion of Gaza. Microsoft has fired five employees following protests at its headquarters in Redmond, according to a statement from the group No Azure for Apartheid.

Microsoft President Brad Smith, whose office the protesters entered, addressed the issue on Thursday. He said that he and Coleman met with Jewish Microsoft employees, who have been harassed and threatened and have seen their public information shared online.

“We don’t get to control what happens outside Microsoft, but we need to be clear about one thing,” Smith said. “There is no room for antisemitism at Microsoft, and as a company and as a community, we will protect this group and defend them from that.”

WATCH: Nebius Co-Founder: $17.4B Microsoft deal highlights surging AI infrastructure demand

Nebius Co-Founder: $17.4 B Microsoft deal highlights surging AI infrastructure demand

Continue Reading

Trending