For a company worth nearly $3 trillion, facing an unexpected cost of a few billion dollars may sound relatively paltry. But U.S. chipmaker Nvidia’s announcement in a regulatory filing earlier this month that it expected to incur costs of up to $5.5 billion as a result of new U.S. export controls sent the company’s stock tumbling more than 6 percent the following day and caused a collective shiver throughout the semiconductor chip industry.
Nvidia’s hefty financial hit comes from a new Trump administration rule requiring the company to acquire a special license to sell its H20 chips in China, adding another hurdle in accessing one of the world’s biggest tech markets and the United States’ foremost competitor in the race for artificial intelligence.
The Trump administration has said that the new license requirement is intended to prevent the chips from being “used in, or diverted to, a supercomputer in China,” according to Nvidia’s filing. It’s the latest attempt by the United States to slow China’s AI development and preserve the United States’ advantage.
The perennial question hanging over U.S. restrictions on Chinese tech over the past eight years has been how well they are actually working. Significant milestones in China this year—such as the launch of the advanced AI model DeepSeek-R1 and advances in semiconductor chips from tech giant Huawei—have reignited that debate.
Some experts and policymakers are now questioning whether it’s too late to keep China from catching up to U.S. AI technology, and whether the United States should instead pursue a more collaborative approach with Beijing on AI development and regulation.
Jensen Huang, the head of the chip company Nvidia, displays some of the company’s graphics cards at the developer conference GTC in San Jose, California, on March 18.Andrej Sokolow/dpa/Reuters
Nvidia created the H20 as a workaround for U.S. government restrictions on another one of its chips—the H800, which the Biden administration banned the company from selling to China in October 2023. The H800 had also been created in response to earlier restrictions by the Biden administration on Nvidia’s sales.
Now, Trump has moved the goalposts again.
“The first round of chip controls came and they set this bar, and then Nvidia said: ‘OK, we’ll build the fanciest thing we can that’s allowed, and sell a bunch of them, because we’ve just been told we could sell those’—and then a bunch of people in Washington were angry, as if this was a sort of unpatriotic thing to do,” said Graham Webster, a research scholar at Stanford University who focuses on China’s tech policy. “I think [Nvidia’s] orientation is pretty consistent—build increasingly advanced chips and sell as many as they can to whoever they can get them to,” he said.
Nvidia’s graphics processing units (GPUs)—a type of semiconductor circuit that the company invented in 1999—have exploded in popularity recently because of their critical role in training artificial intelligence models such as OpenAI’s ChatGPT and its competitors. They have also made the company’s products a prime target of export controls by successive U.S. administrations intent on curbing China’s access to advanced technology.
Trump’s first administration began that effort, restricting Huawei from access to semiconductor chips and other U.S. technology by placing the company (and other Chinese firms) on the Commerce Department’s so-called “entity list” in 2019. The Biden administration broadened the fight in 2022, imposing export restrictions on chips and chipmaking technology to China and continuing to periodically expand those restrictions all the way up until the end of Biden’s term in January.
One of the big uncertainties hanging over Trump’s return to the White House was how his past hawkishness on China pre-Biden would manifest itself post-Biden. While there have been some reversals (see: TikTok) and some continuations (see: trade war), early signs of his second-term strategy to curb China’s semiconductor industry point to more of the same.
But this time around, Trump is facing a slew of recent reminders from China of its continued—and, to Washington, alarming—progress.
A trader reacts on the floor of the New York Stock Exchange at the opening bell in New York City on April 16 as global markets mostly retreated after the U.S. government imposed restrictions on exports of a key Nvidia chip to China. Charly Triballeau/AFP via Getty Images
None of those reminders have been sharper than DeepSeek, whose R1 language model—released in late January—showcased capabilities rivaling those of U.S. leader OpenAI but at a fraction of the cost and computing power.
DeepSeek’s debut sent shock waves through Washington, though experts still debate the extent to which it actually constituted a dreaded “Sputnik moment” for American AI.
“The strength of the reaction in Washington showed that many people didn’t realize what a fast follower China was,” said Helen Toner, the director of strategy and foundational research grants at Georgetown University’s Center for Security and Emerging Technology. “It was a good reality check.”
More pointedly, DeepSeek’s unveiling raised questions for U.S. policymakers about the effectiveness of export controls. That’s because DeepSeek’s success came on the back of American chips—the company trained its model largely using Nvidia’s H800 and H20 GPUs. These chips were acquired legally. DeepSeek stockpiled enough H800s before the Biden administration clamped down on the chips in 2023. In the ever-expanding game of whack-a-mole, the U.S. government was swinging a beat too late.
At the same time, Chinese AI companies’ inability to access the most cutting-edge U.S. chips may have paradoxically supercharged their innovation by forcing them to be more resourceful, as was the case with DeepSeek.
“Here in China, DeepSeek has really encouraged people who pay attention to AI development, because they believe it shows that even under sanctions and different kinds of embargoes of the United States, a Chinese company can still find a way to catch up,” said Xiao Qian, the vice dean of the Institute for AI International Governance and the deputy director of the Center for International Security and Strategy at Tsinghua University in Beijing.
And DeepSeek isn’t alone. China’s top AI models are rapidly closing in on their U.S. peers despite the restrictions, according to benchmarking by Stanford University’s Institute for Human-Centered Artificial Intelligence. In March, Kai-Fu Lee, the Beijing-based CEO of the investment firm 01.AI and a leading AI expert, told Reuters that Chinese AI companies now lag behind U.S. firms by only three months in core technologies.
Chinese tech giants have also been racing to pump out their own advanced chips. Reuters reported that Huawei is preparing to launch its new Ascend 910c AI chip, which Chinese companies are expected to use to replace H20s, as soon as next month. The company is also testing the 910d, which it hopes will supersede the power of Nvidia’s H100—one of the previously banned chips—for model training.
“There’s a very strong sense of insecurity here in China,” Xiao said. “Because of the unpredictability of the Trump administration, we really don’t know what is ahead, so it is natural that all the companies within China are trying to be more self-sufficient, even though at the moment they are still very strongly dependent on the Nvidia chips.”
Taken together, China’s advancements haven’t exactly been a glowing testament to U.S. export restrictions. Yet many experts argue that the policy still has legs.
Chinese AI companies have continued to do whatever they can to buy U.S. chips, which proves their superior quality and performance, said Miles Brundage, a nonresident senior fellow at the Institute for Progress who previously worked as the head of policy research at OpenAI. Before Trump brought the hammer down on H20 chips, Chinese companies had placed orders for 1.3 million of them, totaling more than $16 billion.
The H20s were highly sought after because they are specifically designed for inference—the actual running of a trained model, which is becoming an increasing focus of AI innovation as AI is used more widely. Depriving Chinese companies of these chips could present a real stumbling block.
“In terms of setting back the kind of scale of near-term AI training runs, as well as inference, perhaps more importantly, in China, I’d say it’s a big deal,” Brundage said.
And for the next round of AI advancements, some experts argue that the sheer volume of advanced chips is still a difference maker.
China “finds a lot of ways to come up with innovative developments that maybe are less compute intensive, but they still haven’t quite worked around the fact that the U.S. is still the lead in compute, and we still have access to the most chips and the most computing resources, and that scale still really matters,” said Liza Tobin, the managing director at the geopolitical risk advisory firm Garnaut Global, who previously served as the China director for the National Security Council under both the Biden and Trump administrations. “The demands of scaling and AI just keep going up and up, and that still does give the U.S. an advantage, but it’s not an absolute advantage, and it’s not a permanent advantage.”
People stand in front of a stealth combat drone on display at an exhibition entitled “Forging Ahead in the New Era” in Beijing on Oct. 12, 2022. Noel Celis/AFP via Getty Images
Even though the U.S. policy of restricting China likely has an expiration date, proponents argue that it is worth pursuing as long as possible for one reason above all else: the military implications. Both the Trump and Biden administrations have pointed to the potential for AI to confer new military advantages to China as the primary driver of U.S. policy.
There is still debate about how significantly AI could supercharge China’s military capabilities. The opacity of the People’s Liberation Army has made it hard for researchers to assess China’s progress and plans. Experts describe a wide range of concerns, ranging from the more mundane—including AI models being applied to increase supply-chain efficiency for ammunition and other battlefield resources—to the more nightmarish, such as AI being used to control vast swarms of drones in an invasion of Taiwan.
Due to China’s military-civil fusion policy, which calls for harnessing cutting-edge commercial technologies to strengthen the military, advocates for U.S. controls say that it is necessary for the United States to continue targeting the flow of advanced chips to China as a whole.
“Slowing down China’s military modernization is so important that we should take some risks and incur some costs, especially in those areas where China might be using our capabilities … particularly these high-end chips,” said Jacob Stokes, the deputy director of the Indo-Pacific Security Program at the Center for a New American Security.
But another camp argues that the costs are too high—especially considering that the policy will likely only buy the United States a limited amount of time.
The most visible cost is the hit to corporations’ bottom lines. Nvidia, with a market capitalization not far behind the United Kingdom’s GDP, certainly isn’t the most sympathetic victim, and some AI scholars have argued that soaring demand for the company’s chips in Western nations means that it can easily compensate for the lost revenue from the China market.
But others warn that Washington’s restrictions will eventually come at a cost for U.S. companies, which will be increasingly cut out of the Chinese market as its ecosystem becomes more independent. If U.S. companies do see an overall hit to their revenue, that could reduce their research and development budgets and cause them to lose ground to Chinese competitors over time.
“The restrictions on the H20 are a particularly egregious example of the ‘small gain, high cost’ policy the U.S. has pursued with respect to U.S. hardware companies and China,” said Paul Triolo, a partner at the advisory firm DGA-Albright Stonebridge Group who leads the firm’s technology practice.
Of greatest concern to Triolo and others who question the logic of export restrictions is that AI safety conversations have been displaced by the all-out effort to win the U.S.-China race. During the Biden administration, there were efforts to simultaneously curb China and collaborate on safety standards, with some success. In the final months of the administration, both sides agreed to maintain human control over nuclear weapons.
For these critics, the U.S. focus on restrictions is undermining further safety talks.
“The international discussion on this is very, very limited, and because China and the U.S. lack trust, it is impossible for the two countries to talk about this at the moment,” Xiao said. “We now rely on each country to be self-disciplined, but that is really not a way forward.”
Even for those who support continued restrictions, Washington’s lack of plan for a future of AI parity with China is a concern.
“I think clearly on net it is good to restrict the supply” of chips in the near term, Brundage said. But, he added, the United States also needs to “plan ahead for a scenario where we’ll have to eventually work together on shared safety and security standards and prepare for the kind of military and other consequences of China making these advances in AI. I think it’s good to delay them. But delaying is not a long-term solution.”
For now, the Trump team has indicated that those discussions are not a priority.
“The AI future is not going to be won by hand-wringing about safety,” U.S. Vice President J.D. Vance said in a speech at the AI Action Summit in Paris in February.
The next big test for the United States’ ambitions to outpace China will be the extent to which it can bring the rest of the world—particularly traditional U.S. allies and partners—into the fold.
The imminent challenge for Trump on that front is finalizing another Biden holdover. The Biden administration pushed its “Framework for Artificial Intelligence Diffusion” out the door a week before Trump’s return to the White House.
The framework, more commonly known as the AI diffusion rule, divides countries into three tiers of access to advanced U.S. AI technology. The top tier features 18 close U.S. allies who enjoy near-unrestricted access, including Canada, Germany, and Taiwan, while the bottom tier includes roughly two dozen arms-embargoed adversaries such as China, Russia, North Korea, and Iran, where chips exports are totally banned.
Most other (more than 150) countries are in the second tier, which will be subject to strict licensing requirements for advanced chips and software parameters critical to developing AI models and data centers.
The Biden administration included a 120-day public comment period that kicked the rule’s implementation down the road into the hands of the Trump administration, setting May 15 as the deadline for countries and companies—many of which, including Nvidia, Microsoft, the United Arab Emirates, and the Czech Republic, have lobbied against it—to comply. The final shape that rule takes under Trump will be seen as a bellwether for U.S. AI strategy going forward.
The Trump administration has thus far provided few windows into its thinking, with the closest and most recent coming during a confirmation hearing for Jeffrey Kessler, the Commerce Department’s new undersecretary for industry and security, who will oversee export control implementation. Kessler described the AI diffusion rule as “very complex and bureaucratic,” adding that it was “one of the things I would like to review” once confirmed.
“The identification of the problem was largely correct, but I am not sure this is the right solution,” he said.
That Trump instinct to try to simplify policy (the administration is reportedly considering scrapping the country tiers altogether) as much as possible could run counter to the president’s broader China containment strategy, said Toner of Georgetown University.
“It can be simple, or you can constrain China, or you can help U.S. industry,” she said. “You have to pick two of those three.”
