Connect with us

Published

on

China is focusing on large language models (LLMs) in the artificial intelligence space. 

Blackdovfx | Istock | Getty Images

China’s attempts to dominate the world of artificial intelligence could be paying off, with industry insiders and technology analysts telling CNBC that Chinese AI models are already hugely popular and are keeping pace with — and even surpassing — those from the U.S. in terms of performance.

AI has become the latest battleground between the U.S. and China, with both sides considering it a strategic technology. Washington continues to restrict China’s access to leading-edge chips designed to help power artificial intelligence amid fears that the technology could threaten U.S. national security.

It’s led China to pursue its own approach to boosting the appeal and performance of its AI models, including relying on open-sourcing technology and developing its own super-fast software and chips.

China is creating popular LLMs

On Hugging Face, a repository of LLMs, Chinese LLMs are the most downloaded, according to Tiezhen Wang, a machine learning engineer at the company. Qwen, a family of AI models created by Chinese e-commerce giant Alibaba, is the most popular on Hugging Face, he said.

“Qwen is rapidly gaining popularity due to its outstanding performance on competitive benchmarks,” Wang told CNBC by email.

He added that Qwen has a “highly favorable licensing model” which means it can be used by companies without the need for “extensive legal reviews.”

Qwen comes in various sizes, or parameters, as they’re known in the world of LLMs. Large parameter models are more powerful but have higher computational costs, while smaller ones are cheaper to run.

“Regardless of the size you choose, Qwen is likely to be one of the best-performing models available right now,” Wang added.

DeepSeek, a start-up, also made waves recently with a model called DeepSeek-R1. DeepSeek said last month that its R1 model competes with OpenAI’s o1 — a model designed for reasoning or solving more complex tasks.

These companies claim that their models can compete with other open-source offerings like Meta‘s Llama, as well as closed LLMs such as those from OpenAI, across various functions.

“In the last year, we’ve seen the rise of open source Chinese contributions to AI with really strong performance, low cost to serve and high throughput,” Grace Isford, a partner at Lux Capital, told CNBC by email.

China pushes open source to go global

Open sourcing a technology serves a number of purposes, including driving innovation as more developers have access to it, as well as building a community around a product.

It is not only Chinese firms that have launched open-source LLMs. Facebook parent Meta, as well as European start-up Mistral, also have open-source versions of AI models.

But with the technology industry caught in the crosshairs of the geopolitical battle between Washington and Beijing, open-source LLMs give Chinese firms another advantage: enabling their models to be used globally.

“Chinese companies would like to see their models used outside of China, so this is definitively a way for companies to become global players in the AI space,” Paul Triolo, a partner at global advisory firm DGA Group, told CNBC by email.

While the focus is on AI models right now, there is also debate over what applications will be built on top of them — and who will dominate this global internet landscape going forward.

“If you assume these frontier base AI models are table stakes, it’s about what these models are used for, like accelerating frontier science and engineering technology,” Lux Capital’s Isford said.

Today’s AI models have been compared to operating systems, such as Microsoft’s Windows, Google‘s Android and Apple‘s iOS, with the potential to dominate a market, like these companies do on mobile and PCs.

If true, this makes the stakes for building a dominant LLM higher.

“They [Chinese companies] perceive LLMs as the center of future tech ecosystems,” Xin Sun, senior lecturer in Chinese and East Asian business at King’s College London, told CNBC by email.

“Their future business models will rely on developers joining their ecosystems, developing new applications based on the LLMs, and attracting users and data from which profits can be generated subsequently through various means, including but far beyond directing users to use their cloud services,” Sun added.

Chip restrictions cast doubt over China’s AI future

AI models are trained on vast amounts of data, requiring huge amounts of computing power. Currently, Nvidia is the leading designer of the chips required for this, known as graphics processing units (GPUs).

Most of the leading AI companies are training their systems on Nvidia’s most high-performance chips — but not in China.

Over the past year or so, the U.S. has ramped up export restrictions on advanced semiconductor and chipmaking equipment to China. It means Nvidia‘s leading-edge chips cannot be exported to the country and the company has had to create sanction-compliant semiconductors to export.

Despite, these curbs, however, Chinese firms have still managed to launch advanced AI models.

“Major Chinese technology platforms currently have sufficient access to computing power to continue to improve models. This is because they have stockpiled large numbers of Nvidia GPUs and are also leveraging domestic GPUs from Huawei and other firms,” DGA Group’s Triolo said.

Indeed, Chinese companies have been boosting efforts to create viable alternatives to Nvidia. Huawei has been one of the leading players in pursuit of this goal in China, while firms like Baidu and Alibaba have also been investing in semiconductor design.

“However, the gap in terms of advanced hardware compute will become greater over time, particularly next year as Nvidia rolls out its Blackwell-based systems that are restricted for export to China,” Triolo said.

Lux Capital’s Isford flagged that China has been “systematically investing and growing their whole domestic AI infrastructure stack outside of Nvidia with high-performance AI chips from companies like Baidu.”

“Whether or not Nvidia chips are banned in China will not prevent China from investing and building their own infrastructure to build and train AI models,” she added.

Continue Reading

Technology

OpenAI wins $200 million U.S. defense contract

Published

on

By

OpenAI wins 0 million U.S. defense contract

OpenAI CEO Sam Altman speaks during the Snowflake Summit in San Francisco on June 2, 2025.

Justin Sullivan | Getty Images News | Getty Images

OpenAI has been awarded a $200 million contract to provide the U.S. Defense Department with artificial intelligence tools.

The department announced the one-year contract on Monday, months after OpenAI said it would collaborate with defense technology startup Anduril to deploy advanced AI systems for “national security missions.”

“Under this award, the performer will develop prototype frontier AI capabilities to address critical national security challenges in both warfighting and enterprise domains,” the Defense Department said. It’s the first contract with OpenAI listed on the Department of Defense’s website.

Anduril received a $100 million defense contract in December. Weeks earlier, OpenAI rival Anthropic said it would work with Palantir and Amazon to supply its AI models to U.S. defense and intelligence agencies.

Sam Altman, OpenAI’s co-founder and CEO, said in a discussion with OpenAI board member and former National Security Agency leader Paul Nakasone at a Vanderbilt University event in April that “we have to and are proud to and really want to engage in national security areas.”

OpenAI did not immediately respond to a request for comment.

The Defense Department specified that the contract is with OpenAI Public Sector LLC, and that the work will mostly occur in the National Capital Region, which encompasses Washington, D.C., and several nearby counties in Maryland and Virginia.

Meanwhile, OpenAI is working to build additional computing power in the U.S. In January, Altman appeared alongside President Donald Trump at the White House to announce the $500 billion Stargate project to build AI infrastructure in the U.S.

The new contract will represent a small portion of revenue at OpenAI, which is generating over $10 billion in annualized sales. In March, the company announced a $40 billion financing round at a $300 billion valuation.

In April, Microsoft, which supplies cloud infrastructure to OpenAI, said the U.S. Defense Information Systems Agency has authorized the use of the Azure OpenAI service with secret classified information. 

WATCH: OpenAI hits $10 billion in annual recurring revenue

OpenAI hits $10 billion in annual recurring revenue

Continue Reading

Technology

Amazon Kuiper second satellite launch postponed by ULA due to rocket booster issue

Published

on

By

Amazon Kuiper second satellite launch postponed by ULA due to rocket booster issue

A United Launch Alliance Atlas V rocket is shown on its launch pad carrying Amazon’s Project Kuiper internet network satellites as the vehicle is prepared for launch at the Cape Canaveral Space Force Station in Cape Canaveral, Florida, U.S., April 28, 2025.

Steve Nesius | Reuters

United Launch Alliance on Monday was forced to delay the second flight carrying a batch of Amazon‘s Project Kuiper internet satellites because of a problem with the rocket booster.

With roughly 30 minutes left in the countdown, ULA announced it was scrubbing the launch due to an issue with “an elevated purge temperature” within its Atlas V rocket’s booster engine. The company said it will provide a new launch date at a later point.

“Possible issue with a GN2 purge line that cannot be resolved inside the count,” ULA CEO Tory Bruno said in a post on Bluesky. “We will need to stand down for today. We’ll sort it and be back.”

The launch from Florida’s Space Coast had been set for last Friday, but was rescheduled to Monday at 1:25 p.m. ET due to inclement weather.

Read more CNBC tech news

Amazon in April successfully sent up 27 Kuiper internet satellites into low Earth orbit, a region of space that’s within 1,200 miles of the Earth’s surface. The second voyage will send “another 27 satellites into orbit, bringing our total constellation size to 54 satellites,” Amazon said in a blog post.

Kuiper is the latest entrant in the burgeoning satellite internet industry, which aims to beam high-speed internet to the ground from orbit. The industry is currently dominated by Elon Musk’s Space X, which operates Starlink. Other competitors include SoftBank-backed OneWeb and Viasat.

Amazon is targeting a constellation of more than 3,000 satellites. The company has to meet a Federal Communications Commission deadline to launch half of its total constellation, or 1,618 satellites, by July 2026.

Don’t miss these insights from CNBC PRO

AWS CEO: Lots of opportunity to expand infrastructure globally

Continue Reading

Technology

Google issues apology, incident report for hourslong cloud outage

Published

on

By

Google issues apology, incident report for hourslong cloud outage

Thomas Kurian, CEO of Google Cloud, speaks at a cloud computing conference held by the company in 2019.

Michael Short | Bloomberg | Getty Images

Google apologized for a major outage that the company said was caused by multiple layers of flawed recent updates.

The company released an incident report late on Friday that explained hours of downtime on Thursday. More than 70 Google cloud services stopped working properly across the globe, knocking down or disrupting dozens of third-party services, including Cloudflare, OpenAI and Shopify. Gmail, Google Calendar, Google Drive, Google Meet and other first-party products also malfunctioned.

“We deeply apologize for the impact this outage has had,” Google wrote in the incident report. “Google Cloud customers and their users trust their businesses to Google, and we will do better. We apologize for the impact this has had not only on our customers’ businesses and their users but also on the trust of our systems. We are committed to making improvements to help avoid outages like this moving forward.”

Thomas Kurian, CEO of Google’s cloud unit, also posted about the outage in an X post on Thursday, saying “we regret the disruption this caused our customers.”

Google in May added a new feature to its “quota policy checks” for evaluating automated incoming requests, but the new feature wasn’t immediately tested in real-world situations, the company wrote in the incident report. As a result, the company’s systems didn’t know how to properly handle data from the new feature, which included blank entries. Those blank entries were then sent out to all Google Cloud data center regions, which prompted the crashes, the company wrote.

Engineers figured out the issue in 10 minutes, according to the company. However, the entire incident went on for seven hours after that, with the crash leading to an overload in some larger regions.

As it released the feature, Google did not use feature flags, an increasingly common industry practice that allows for slow implementation to minimize impact if problems occur. Feature flags would have caught the issue before the feature became widely available, Google said.

Going forward, Google will change its architecture so if one system fails, it can still operate without crashing, the company said. Google said it will also audit all systems and improve its communications “both automated and human, so our customers get the information they need asap to react to issues.” 

— CNBC’s Jordan Novet contributed to this report.

WATCH: Google buyouts highlight tech’s cost-cutting amid AI CapEx boom

Google buyouts highlight tech's cost-cutting amid AI CapEx boom

Continue Reading

Trending