China is looking to challenge the U.S. in artificial intelligence. China’s tech giants have launched their own AI models.
Niphon | Istock | Getty Images
China’s race to develop smarter-than-human artificial intelligence may put it ahead of the U.S., but such ground-breaking technology could also risk lessening the stronghold that the ruling Communist Party has over the world’s second-largest economy.
That’s the view of prominent AI scientist Max Tegmark, who told CNBC artificial general intelligence (AGI) is closer than we think and the narrative of a geopolitical battle between the U.S. and China racing to build the smartest AI is a “suicide race.”
While there is no singular definition of AGI, it is broadly taken to refer to AI that can outsmart humans.
Applications like ChatGPT — that allow users to prompt a chatbot for answers — have exploded in popularity. But many AI companies are racing to develop the next level, with AI that has human-level intelligence.
Sam Altman, the CEO of OpenAI, has said that AGI could be achieved by 2025. While there are other major names in the tech sector who also think AGI is close, many others think true AGI is still very far away.
As well as competition between technology companies, there is also the geopolitical battle taking place between the U.S. and China for dominance in realms from AI to chips. While this is often portrayed as a race to be first to the latest technology, Tegmark said this is not the right framing.
“I think of this battle, this geopolitical battle to build AGI first as a ‘hopium war’,” Tegmark told CNBC in an interview last month. ” I call it the ‘hopium war’ because it’s fueled by … delusional hope that we can control AGI.”
Tegmark is the president of the Future of Life Institute, a thinktank which penned a letter last year calling for AI labs to pause the development of advanced AI systems. The letter was signed by major tech names including Tesla CEO Elon Musk. Tegmark’s concern is that AI is advancing rapidly with very few guardrails in place, and no way to control it should it begin to outsmart humans.
“We are much closer to building AGI than figuring out how to control it. And that means that the AGI race is not an arms race, it’s a suicide race,” Tegmark said.
Is China worried about AGI?
China has little incentive to build AGI, according to Tegmark. The AI scientist recalled a story in which Elon Musk told him about a “high level meeting” the Tesla boss had with Chinese government officials in early 2023. Musk said to the Chinese government that if AGI is built, China “will not be controlled by the Communist Party, but by the super intelligence,” Tegmark said.
“[Musk] got a very strong reaction. Some of them, really hadn’t thought about that, and with less than a month from that, China came out with their first AI regulations,” Tegmark said, referencing new regulation governing generative AI.
China’s ministry of foreign affairs was not immediately available for comment on the anecdote. CNBC also contacted Tesla for a response from Musk.
“The U.S. doesn’t need to convince China to not build AGI. Even if the U.S. didn’t exist, the Chinese government would have an incentive to not build it because they want to be in control,” Tegmark said.
“[The] last thing they want is to lose that control.”
China’s approach to AI
AI is a strategic priority for the Chinese government. The country’s biggest firms such as Alibaba, Huawei and Tencent have been developing their own AI models. The capabilities of those models are also advancing.
China was also among the first countries in the world to bring in regulation around various aspects of AI. The country’s internet is heavily censored and any information that appears to go against Beijing’s ideology is blocked. OpenAI’s ChatGPT is banned and it is well-noted that chatbots in China won’t answer questions related to politics and topics deemed sensitive by the Communist Party.
The country’s approach to AI is therefore an attempt to push innovation while also balancing its own interests. When it comes to AGI, China is likely to pursue a similar approach, according to analysts.
“I would not count on China to limit its own AI capabilities due to fears that such technologies would threaten Party rule. Similar predictions were made about the internet, they all proved to be false,” Kendra Schaefer, a partner at consultancy Trivium China,” told CNBC.
“China will attempt to dominate AGI while creating a techno-regulatory apparatus that limits what AGI is permitted to do domestically.”
U.S.-China AI battle
Despite Tegmark’s view that the the race to build AGI is a “hopium war,” geopoltiics remains front and center between the U.S. and China when it comes to development of the technology.
“Right now, China is viewing AI through a dual-lens: geopolitical power and domestic growth,” said Abishur Prakash, founder and geopolitical strategist at Toronto-based strategy advisory firm, The Geopolitical Business.
“With AI, China hopes to shift the balance of power around the globe, like creating a new export model. And, in parallel, China wants to power its economy in new ways, from government efficiency to business applications,” Prakash told CNBC.
The U.S. has pursued a policy of attempting to restrict China’s access to key technologies, mainly semiconductors like those designed by Nvidia, that are required to train more advanced AI models. China has responded by attempting to build its homegrown chip industry.
Will the U.S. and China partner on AI rules?
Technologists have warned of some of the risks and dangers when AGI does finally arrive. One theory is that without guardrails, AI will be able to improve itself and design new systems independently.
Tegmark believes that any such risks will be realized by both the U.S. and China, which will force both countries’ governments to individually come up with rules around AI safety.
“So my optimistic path forward is the U.S. and China unilaterally impose national safety standards to prevent their own companies from doing harm and building uncontrollable AGI, not to appease the rivals superpowers, but just to protect themselves,” Tegmark said.
“After that happens though, there’s this really interesting stage where the U.S. and China will be like, wait, how can we guarantee that North Korea doesn’t build AGI or someone else? And then the U.S. and China have an incentive now to push the rest of the world to join them into an AGI moratorium.”
Indeed, governments are already trying to work together to figure out how to create regulations and frameworks around AI. Last year, the U.K. hosted an AI safety summit, which the U.S. and China were both in attendance, to discuss potential guardrails around the technology.
But regulation and rules around AI are currently fragmented. This year, the European Union enacted the AI Act, the first major law globally governing the technology. China has its own set of rules, while many other countries have not yet moved to create any regulation.
Tegmark’s hope of co-ordination around AI safety is echoed by others.
“When the dangers of competition are greater than the rewards, nations will ideally be motivated to come together and mutually self-regulate,” Trivium China’s Schaefer said.
“Indeed, some Chinese policymakers have advocated for getting out ahead of that potential issue and establishing an international governance body under the UN – similar to the International Atomic Energy Agency – so there is desire on Beijing’s side to establish a global governance body,” she said.
Silicon Valley executives and financiers publicly opened their wallets in support of President Donald Trump’s 2024 presidential run. The early returns in 2025 aren’t great, to say the least.
Following Trump’s sweeping tariff plan announced Wednesday, the Nasdaq suffered steep consecutive daily drops to finish 10% lower for the week, the index’s worst performance since the beginning of the Covid pandemic in 2020.
The tech industry’s leading CEO’s rushed to contribute to Trump’s inauguration in January and paraded to Washington, D.C., for the event. Since then, it’s been a slog.
The market can always turn around, but economists and investors aren’t optimistic, and concerns are building of a potential recession. The seven most valuable U.S. tech companies lost a combined $1.8 trillion in market cap in two days.
Apple slid 14% for the week, its biggest drop in more than five years. Tesla, led by top Trump adviser Elon Musk, plunged 9.2% and is now down more than 40% for the year. Musk contributed close to $300 million to help propel Trump back to the White House.
Nvidia, Meta and Amazon all suffered double-digit drops for the week. For Amazon, a ninth straight weekly decline marks its longest such losing streak since 2008.
With Wall Street selling out of risky assets on concern that widespread tariff hikes will punish the U.S. and global economy, the fallout has drifted down to the IPO market. Online lender Klarna and ticketing marketplace StubHub delayed their IPOs due to market turbulence, just weeks after filing with the Securities and Exchange Commission, and fintech company Chime is also reportedly delaying its listing.
CoreWeave, a provider of artificial intelligence infrastructure, last week became the first venture-backed company to raise more than $1 billion in a U.S. IPO since 2021. But the company slashed its offering, and trading has been very volatile in its opening days on the market. The stock plunged 12% on Friday, leaving it 17% above its offer price but below the bottom of its initial range.
“You couldn’t create a worse market and macro environment to go public,” said Phil Haslett, co-founder of EquityZen, a platform for investing in private companies. “Way too much turbulence. All flights are grounded until further notice.”
CoreWeave investor Mark Klein of SuRo Capital previously told CNBC that the company could be the first in an “IPO parade.” Now he’s backtracking.
“It appears that the IPO parade has been temporarily halted,” Klein told CNBC by email on Friday. “The current tariff situation has prompted these companies to pause and assess its impact.”
‘Cave rapidly’
During last year’s presidential campaign, prominent venture capitalists like Marc Andreessen backed Trump, expecting that his administration would usher in a boom and eliminate some of the hurdles to startup growth set up by the Biden administration. Andreessen and his partner, Ben Horowitz, said in July that their financial support of the Trump campaign was due to what they called a better “little tech agenda.”
A spokesperson for Andreessen Horowitz declined to comment.
Some techies who supported Trump in the campaign have taken to social media to defend their positions.
Venture capitalist Keith Rabois, a managing director at Khosla Ventures, posted on X on Thursday that “Trump Derangement Syndrome has morphed into Tariff Derangement Syndrome.” He said tariffs aren’t inflationary, are effective at reducing fentanyl imports, and he expects that “most other countries will cave and cave rapidly.”
That was before China’s Finance Ministry said on Friday that it will impose a 34% tariff on all goods imported from the U.S. starting on April 10.
At Sequoia Capital, which is the biggest investor in Klarna, outspoken Trump supporter Shaun Maguire, wrote on X, “The first long-term thinking President of my lifetime,” and said in a separate post that, “The price of stocks says almost nothing about the long term health of an economy.”
However, Allianz Chief Economic Advisor Mohamed El-Erian warned on Friday that Trump’s extensive raft of import tariffs are putting the U.S. economy at risk of recession.
“You’ve had a major repricing of growth prospects, with a recession in the U.S. going up to 50% probability, you’ve seen an increase in inflation expectations, up to 3.5%,” he told CNBC’s Silvia Amaro on the sidelines of the Ambrosetti Forum in Cernobbio, Italy.
Former Microsoft CEOs Bill Gates, left, and Steve Ballmer, center, pose for photos with CEO Satya Nadella during an event celebrating the 50th Anniversary of Microsoft on April 4, 2025 in Redmond, Washington.
Stephen Brashear | Getty Images
Meanwhile, executives at tech’s megacap companies were largely silent this week, and their public relations representatives declined to provide comments about their thinking.
Microsoft CEO Satya Nadella was in the awkward position on Friday of celebrating his company’s 50th anniversary at corporate headquarters in Redmond, Washington. Alongside Microsoft’s prior two CEOs, Bill Gates and Steve Ballmer, Nadella sat down with CNBC’s Andrew Ross Sorkin for a televised interview that was planned well before Trump’s tariff announcement.
When asked about the tariffs at the top of the interview, Nadella effectively dodged the question and avoided expressing his views about whether the new policies will hamper Microsoft’s business.
Ballmer, who was succeeded by Nadella in 2014, acknowledged to Sorkin that “disruption is very hard on people” and that, “as a Microsoft shareholder, this kind of thing is not good.” Ballmer and Gates are two of the 12 wealthiest people in the world thanks to their Microsoft fortunes.
C-suites may not be able to stay quiet for long, especially if the recent turmoil spills into next week.
Lise Buyer, who previously helped guide Google through its IPO and now works as an adviser to companies going public, said there’s no appetite for risk in the market under these conditions. But there is risk that staffers get jittery, and they’ll surely look to their leaders for some reassurance.
“Until markets settle out and we have the opportunity to access valuation levels, public company CEOs should work to calm potentially distressed employees,” Buyer said in an email. “And private company managements should refine plans to get by on dollars already in the treasury.”
— CNBC’s Hayden Field, Jordan Novet, Leslie Picker, Annie Palmer and Samantha Subin contributed to this report.
Elon Musk has been promising investors for about a decade that Tesla’s cars are on the verge of turning into robotaxis, capable of driving themselves cross-country, after one big software update.
That hasn’t happened yet.
What Tesla offers is a sophisticated, but only partially automated, driving system that’s marketed in the U.S. as its Full Self-Driving (Supervised) option, though many Tesla fans refer to it as FSD. In China, Tesla recently changed the system’s name to “intelligent assisted driving.”
Full Self-Driving, as it was previously called, relies on cameras and software to enable features like automatic navigation on highways and city streets, or automatic braking and slowing in response to traffic lights and stop signs.
Tesla owner’s manuals warn users that FSD “is a hands-on feature” that requires them to pay attention to the road at all times. “Keep your hands on the steering wheel at all times, be mindful of road conditions and surrounding traffic,” the manuals say.
But many of Tesla’s customers ignore the fine print and use the system hands-free anyway.
Tesla’s partially automated driving systems have been a source of inspiration for its stalwart fans. But they’ve also caused controversy and concern for public safety after reports of injurious and fatal collisions where Tesla’s standard Autopilot or premium FSD systems were known to be in use.
FSD does a lot of things “amazingly well,” said Guy Mangiamele, a professional test driver for automotive consulting firm AMCI Testing, during a recent long drive in Los Angeles. But he added that “the times that it trips up, you could kill somebody or you could hurt yourself.”
The pressure has never been higher on Tesla to elevate the technology and deliver on Musk’s long-delayed promises.
The Tesla CEO is the wealthiest person in the world and was the biggest financial backer of President Donald Trump’s 2024 campaign. Since Trump’s January inauguration, Musk has been leading the administration’s Department of Government Efficiency effort to drastically slash the federal workforce and government spending.
The DOGE team has been connected to more than 280,000 layoff plans for federal workers and contractors impacting 27 agencies over the last two months, according to data tracked by Challenger Gray, the executive outplacement firm.
Musk’s work with DOGE – along with his frequently incendiary political rhetoric and endorsement of Germany’s far-right, anti-immigrant party AfD – has led to a tremendous backlash against Tesla.
Protests, boycotts and even criminal acts of vandalism have targeted the electric vehicle maker in recent months and led many prospective Tesla customers to turn to other brands. Meanwhile, existing Tesla owners have been trading in their EVs at record levels, according to data from Edmunds.
Tesla’s stock dropped 36% through the first three months of 2025, representing its steepest decline since 2022 and third-biggest slide for any quarter since the EV maker went public in June 2010. Tesla also reported 336,681 vehicle deliveries in the first quarter of 2025, a 13% decline from the same period a year ago.
Product unveilings and a “robotaxi launch” expected from Tesla in Austin, Texas, this year could revitalize investors’ sentiment about the company and hopefully lift its share price, Piper Sandler analysts wrote in a note following the worse-than-expected deliveries report.
On Tesla’s last earnings call, Musk promised investors that Tesla will finally start its driverless ride-hailing service in Austin in June.
To see whether the company’s FSD technology is anywhere close to a robotaxi-ready release, CNBC spent months riding along with Tesla owners who use Full Self-Driving (Supervised) and speaking with automotive safety experts about their impressions.
Auto-tech enthusiast and Tesla owner Chris Lee, host of the YouTube channel EverydayChris, told CNBC that Tesla’s system “definitely has a ways to go, but the fact that it’s able to go from where it was three years ago to today, is insane.”
Many experts, including Telemetry Vice President of Market Research Sam Abuelsamid, remain skeptical. There’s been “no evidence” that FSD is “anywhere close to being ready to be used in an unsupervised form” by June, said Abuelsamid, whose firms specializes in automotive intelligence.
Tesla FSD will “often work really well, particularly in daytime conditions” but then “randomly, in a scenario where it did fine previously, it will fail,” said Abuelsamid, adding that those scenarios can be unpredictable and dangerous.
Watch the video to learn more about the evolution of Tesla’s Full Self-Driving (Supervised) and whether it will be robotaxi-ready this June.
Microsoft owns lots of Nvidia graphics processing units, but it isn’t using them to develop state-of-the-art artificial intelligence models.
There are good reasons for that position, Mustafa Suleyman, the company’s CEO of AI, told CNBC’s Steve Kovach in an interview on Friday. Waiting to build models that are “three or six months behind” offers several advantages, including lower costs and the ability to concentrate on specific use cases, Suleyman said.
It’s “cheaper to give a specific answer once you’ve waited for the first three or six months for the frontier to go first. We call that off-frontier,” he said. “That’s actually our strategy, is to really play a very tight second, given the capital-intensiveness of these models.”
Suleyman made a name for himself as a co-founder of DeepMind, the AI lab that Google bought in 2014, reportedly for $400 million to $650 million. Suleyman arrived at Microsoft last year alongside other employees of the startup Inflection, where he had been CEO.
More than ever, Microsoft counts on relationships with other companies to grow.
It gets AI models from San Francisco startup OpenAI and supplemental computing power from newly public CoreWeave in New Jersey. Microsoft has repeatedly enriched Bing, Windows and other products with OpenAI’s latest systems for writing human-like language and generating images.
Microsoft’s Copilot will gain “memory” to retain key facts about people who repeatedly use the assistant, Suleyman said Friday at an event in Microsoft’s Redmond, Washington, headquarters to commemorate the company’s 50th birthday. That feature came first to OpenAI’s ChatGPT, which has 500 million weekly users.
Through ChatGPT, people can access top-flight large language models such as the o1 reasoning model that takes time before spitting out an answer. OpenAI introduced that capability in September — only weeks later did Microsoft bring a similar capability called Think Deeper to Copilot.
Microsoft occasionally releases open-source small-language models that can run on PCs. They don’t require powerful server GPUs, making them different from OpenAI’s o1.
OpenAI and Microsoft have held a tight relationship shortly after the startup launched its ChatGPT chatbot in late 2022, effectively kicking off the generative AI race. In total, Microsoft has invested $13.75 billion in the startup, but more recently, fissures in the relationship between the two companies have begun to show.
Microsoft added OpenAI to its list of competitors in July 2024, and OpenAI in January announced that it was working with rival cloud provider Oracle on the $500 billion Stargate project. That came after years of OpenAI exclusively relying on Microsoft’s Azure cloud. Despite OpenAI partnering with Oracle, Microsoft in a blog post announced that the startup had “recently made a new, large Azure commitment.”
“Look, it’s absolutely mission-critical that long-term, we are able to do AI self-sufficiently at Microsoft,” Suleyman said. “At the same time, I think about these things over five and 10 year periods. You know, until 2030 at least, we are deeply partnered with OpenAI, who have [had an] enormously successful relationship for us.
Microsoft is focused on building its own AI internally, but the company is not pushing itself to build the most cutting-edge models, Suleyman said.
“We have an incredibly strong AI team, huge amounts of compute, and it’s very important to us that, you know, maybe we don’t develop the absolute frontier, the best model in the world first,” he said. “That’s very, very expensive to do and unnecessary to cause that duplication.”