How Nvidia Uses Its Soft Power to Influence Customers and Shape Policy

Last month, OpenAI Senior Research Scientist Noam Brown posted a link in a company Slack channel to a news story about OpenAI’s changing stance on advanced chip sales to China and asked a blunt question: “What is our policy on chip exports?” OpenAI policy leaders responded on Slack by citing past company posts, such as a safety blueprint on frontier AI published in June, that supported general chip export controls but did not address China specifically. They also directed staffers to speak to members of the global affairs team working on new export control policies, rather than discussing it on Slack.

Chip restrictions are a delicate topic at the company. Over the past year, OpenAI has avoided taking a public stance on the issue of chip export controls for China because OpenAI leaders, including CEO Sam Altman, have expressed concern that doing so will alienate chipmaker Nvidia and its CEO, Jensen Huang, according to people with knowledge of the company’s strategy. Last year, as Congress turned its attention to export control legislation, a senior OpenAI policy staffer privately said OpenAI was not as vocal on chip export controls as it had been under the Biden administration because it feared antagonizing Huang and Nvidia, according to a staffer on Capitol Hill.

Huang has become the most recognizable face of the AI boom by transforming a small gaming-chip company into the world’s most valuable business, worth nearly $5 trillion. Since the start of the second Trump administration, Huang has had an open line to the president, and they have bonded over a shared belief in the importance of AI to the American economy and global competitiveness. The relationship has bought Huang some policy wins, most notably limited sales of Nvidia’s last-generation chips to China, but far from everything the company wants.

Nvidia’s aggressive use of its soft power to protect its business and influence policy has at times surprised people who have interacted with the company. Government officials sometimes have felt bullied in exchanges with Huang. Officials at political think tanks say Nvidia officials have threatened to cut back on their funding for expressing views the company disagrees with. And Nvidia customers say they tiptoe around certain subjects, as OpenAI has, to avoid displeasing the chip giant.

For example, two executives at cloud companies said fear of angering Nvidia is always at the back of their minds when making decisions. Another cloud executive at a company that uses AMD graphics processing units in its data centers said some customers don’t want the cloud company to advertise their use of AMD chips to avoid backlash from Nvidia. (The executive declined to share the customers’ names for the same reason.)

The chipmaker’s most powerful form of leverage with such customers is how it doles out its GPUs, which are the dominant chips for training and running AI models today. The Information has previously reported that in some instances over the past couple of years, some customers felt that Nvidia prioritized those who bought its full suite of hardware and punished those who did not.

In Washington, Nvidia’s most divisive move has been its push to loosen export controls that prevent it from selling its most advanced chips in China. The effort has alienated China hawks in both political parties and eroded trust in Huang, despite his close relationship with Trump. With Congress gearing up for another battle over export control legislation after Tuesday’s midterm elections, Nvidia’s influence on the Hill could wane as its detractors grow in number.

It didn’t help Huang’s reputation in Washington when The New York Times reported last week that U.S. spy agencies had concluded Huang had a close association with a Chinese businesswoman, Huang Le, whom they believed was a Chinese intelligence asset. The U.S. agencies briefed Trump on their findings, but the president did not act on the information. Trump himself appeared to back away from his unconditional support for Huang, telling reporters on Saturday that he continues to have confidence in Huang unless “we find out something to the contrary.”

Chris McGuire, a senior fellow for China and emerging technologies at the Council on Foreign Relations, said Nvidia has created mistrust among policymakers by exaggerating the capabilities of China’s homegrown AI chips as it seeks looser export controls.

“The fact is, all the data says that China’s AI chips lag far behind Nvidia’s in both quality and quantity—but this is not the story that Nvidia tells,” said McGuire, who was senior adviser on technology to the U.S. Department of State during both the Biden and Trump administrations. “However, as policymakers in the administration and on the Hill have come to realize that Nvidia’s assertions about China’s AI chip capabilities are not fully accurate, it has caused Nvidia to lose credibility in Washington.”

Said Chris Lehane, OpenAI’s chief global affairs officer, in a statement: “As we have made clear, OpenAI continues to support the strengthening of export controls. Strong export controls will both support the build out of AI on democratic rails but also give the US and its allies more time to strengthen the safety standards, oversight and international coordination needed to help make sure people always stay in control of AI.”

“America cannot win the AI race unless America’s AI infrastructure wins worldwide,” said John Rizzo, a spokesperson for Nvidia, in a statement. “If the world’s clouds rely on foreign-made chips, the model and application layers will be foreign-built as well.”

Biden Problems

Huang’s relationship with the last presidential administration couldn’t have been more different from the one he enjoys with the Trump White House.

In 2022, Nvidia was caught off guard when the U.S. government banned sales of the most advanced AI chips to China, which then included Nvidia’s H100, according to multiple government officials and industry figures who interacted with Nvidia at the time. At the time, the chipmaker had a bare-bones government affairs operation. Huang personally launched a vigorous campaign to get the administration to reverse the move, calling then–U.S. Commerce Secretary Gina Raimondo on multiple occasions to make his case, according to people familiar with the calls. According to one Biden official, Nvidia had not yet refined its talking points and was not able to convince administration policymakers to make significant changes.

Within months, Huang had a team reengineer the chips, slowing the rate at which they transferred data to get under the ban’s technical thresholds. It subsequently released them as the H800 and A800. While the new chips weren’t as good as Nvidia’s H100s at communicating with each other, they were still advanced enough to allow Chinese companies to train and run powerful AI systems. That led the Biden administration in October 2023 to announce new restrictions banning sale of the reengineered chips in China, which Nvidia responded to by releasing the even more stripped-down H20, for the Chinese market.

One semiconductor industry executive who worked with Nvidia described the company’s effort to tweak its products and sidestep the ban as giving the Biden administration “the finger” and taking the stance that it did not have to negotiate with the government.

That often adversarial attitude toward some Biden officials came not only from Huang but also from Nvidia’s general counsel, Tim Teter, according to several people who have interacted with him. Those people recalled Teter using unflattering nicknames for government officials, academics and think tank scholars he did not like.

After the initial round of Biden export controls came out, Nvidia began to staff up its team in Washington, hiring former senior Homeland Security official Meg King, senior State Department official Ruth Berry and Semiconductor Industry Association Director Eric Breckenfeld in 2023.

Two former Biden officials said Nvidia’s new team didn’t improve relations with the administration—instead, they would not productively engage on technical issues around export controls. Nvidia executives denied the seriousness of issues such as chip smuggling, despite what the government regarded as compelling evidence to the contrary.

Near the end of the Biden administration, Huang and his team fought another policy, issued in January 2025 and known as the Framework for Artificial Intelligence Diffusion, that further limited exports of advanced chips. One Biden official said the new rules were particularly strict in part because Nvidia did not come to the table and negotiate.

The Nvidia CEO told then–National Security Advisor Jake Sullivan that the moves would embolden China’s chip industry, according to people familiar with the conversations. He also questioned why security professionals were now using technology to provoke a military adversary.

Huang argued that the U.S. rules were hurting Nvidia, when the alternative was that China could stay hooked on Nvidia tech forever. He said the company’s chips were just normal technology, and he didn’t understand why the government was treating them as though they were something more dangerous.

Still, Biden officials found Huang’s interactions offputting: They described him as “pugilistic.” The tenor of his approach was not common among the CEOs they dealt with, former Biden officials said.

Tank Battles

Huang quickly found a new ally when Trump reentered the White House in January 2025.

While Nvidia didn’t have much of a Washington presence during Trump’s first term, according to one former senior Trump official, Huang became a fixture in the White House during Trump’s second term. Huang influenced policy discussions around export controls, especially ahead of May 2025, when the Trump administration announced it was rescinding the Biden-era chip controls. It also greenlit new AI deals and chip exports to several Gulf countries to which Nvidia chip sales had previously been restricted.

As Congress began to debate its own legislation to introduce different forms of chip export controls, Nvidia also ratcheted up its lobbying campaigns across Washington. That included putting pressure on think tank scholars advocating for stricter chip export controls.

In March 2025, Sam Hammond, the AI policy director at the right-leaning Foundation for American Innovation, wrote an article for City Journal arguing that U.S. companies like Nvidia were helping China catch up with the U.S. in AI due to leaky export control policies and lax enforcement of chip smuggling. Hammond said within hours of publication, Nvidia’s government affairs team requested a meeting with him, where they denied the existence of chip smuggling. According to Hammond, they even demanded he produce the serial number of a smuggled chip to demonstrate that smuggling was a real phenomenon.

Hammond crossed Nvidia again in August 2025 when he posted a message on X in response to an article about the Chinese military seeking Nvidia chips for a variety of AI projects: “Nvidia is arming China’s military.” Soon after, Hammond said Nvidia emailed his boss, saying the Foundation for American Innovation wasn’t living up to its pro-innovation values and offering to have their general counsels meet.

It’s not unusual for companies to joust with Washington think tanks that criticize them or advocate for policies contrary to their interests, according to people who work for the organizations. As a bridge between academia and government, think tanks play an important role in shaping policy. While companies often donate to them, many think tanks purport to maintain academic and editorial independence, even if the lines often get blurred.

Nvidia especially has a reputation for being outspoken about think tank output it disagrees with. A former researcher at the Center for a New American Security, a think tank focused on national security and defense policy that receives donations from Nvidia, said Nvidia is not the only organization that pushes back on think tank scholarship it disagrees with. But the researcher noted Nvidia’s increasing sway in Washington, with its government affairs staff growing substantially since the Biden years.

Nvidia tangled with think tanks before Trump’s return to the White House. In 2023, several people familiar with the discussions said Nvidia had tried to use its influence—including funding—to sideline Greg Allen, then director of the AI policy arm of the Center for Strategic and International Studies, an influential think tank focused on national security and geopolitics. Allen staunchly supports chip export controls.

One of those people said Nvidia became a major donor to CSIS and coordinated with the think tank on specific policy reports not labeled as sponsored. CSIS did not ultimately fire Allen; he left the think tank earlier this year. The New York Times earlier reported some details on the incident in 2023. A spokesperson for CSIS said in a statement that Nvidia never asked CSIS to fire Greg Allen, and that the think tank discloses funding that supports specific research reports.

Other think tanks say they’ve been on the receiving end of threats from Nvidia during Trump’s second term.

One fellow at a prominent conservative think tank that received donations from Nvidia said he faced complaints earlier this year over his outspoken public critiques of Nvidia—including claims that the Chinese military was using Nvidia chips. In an internal email thread shared with The Information, leaders at the think tank said Nvidia policy leadership had expressed concerns to the think tank over the “veracity” of the fellow’s public statements and his unwillingness to engage on the issue.

According to the email thread, Nvidia leadership said they were supportive of the work of other scholars there, but renewing Nvidia’s donation to the think tank would be a “tough sell internally.” Ultimately, the fellow pushed back on Nvidia’s allegations internally. At a later meeting, the think tank celebrated the episode as exemplifying the independence of its work, according to someone who was there.

‘Trust Deficit’

On Capitol Hill, Nvidia’s hard-line stances on chip export controls have frayed the company’s relations with many lawmakers and their staff, according to those staffers, lobbyists and other policy experts.

One high-profile incident involved Brian Mast, a Republican congressman from Florida. As chair of the influential House Foreign Affairs Committee, Mast supported export controls, drawing the ire of Nvidia. In January, Mast posted a thread on X sharing an article that accused Nvidia of pushing conservative influencers to attack export controls.

“Jensen Huang, since you’re not on X, I’ll tag your company Nvidia,” Mast wrote. “You and your paid minions are fighting to sell millions of advanced AI chips to Chinese military companies like Alibaba and Tencent. I’m trying to stop that from happening.”

Hill staffers, other government officials and lobbyists said Nvidia has strained relationships with many Congressional offices. According to one Hill staffer, the company has created “a trust deficit” in Washington that would take years to recover from. They pointed to other moderate lawmakers increasingly turning their attention on Nvidia, such as Chris Coons, a centrist Democratic senator from Delaware who has become a vocal critic of Nvidia’s chip exports to China.

More recently, some Congressional staffers and lobbyists say they have noticed a softer touch in Nvidia’s interactions on Capitol Hill. As evidence that Nvidia’s government affairs operation is maturing, they pointed to its June hiring of former Intel lobbyist and Obama administration Commerce Department senior official Bruce Andrews as its new chief external affairs officer.

Staffers and lobbyists said Nvidia has been less aggressive during this year’s Congressional fight over chip export control bills. Some of them attributed the change partly to the fact that David Sacks, a close Nvidia ally, left the White House as its crypto and AI czar in March. Part of the shift can also be chalked up to the fact that the Trump administration cleared Nvidia to sell limited quantities of workhorse H200 chips to Chinese companies last December, though Nvidia has yet to recognize material revenue from the sales due to restrictions by the Chinese government.

Plus, Nvidia’s biggest customers are gaining some leverage—which could also sway the balance of power among AI giants in Washington. While OpenAI may be wary of challenging Nvidia on China export control policy, it has been willing to develop its own competitor chip, Jalapeño, which it claims performs better than Nvidia’s Blackwell GPUs on some benchmarks. Unlike Nvidia, OpenAI has said it will deploy its own chips only in the U.S. and allied countries. An OpenAI spokesperson confirmed that the company will not sell its chip to China.

Indeed, Nvidia’s core business is facing more challengers than ever as many of its U.S. customers enter the red-hot market of AI chips. That, according to one semiconductor executive, is the reason it is so insistent on retaining the ability to export its chips overseas.

But with Congress expected to renew the legislative debate over export controls after Tuesday’s midterm elections, Nvidia’s standing in Washington will face a significant test.

“There are so many ways that policymakers could benefit from Nvidia’s incredible technical expertise and deep market insight to help craft policies that both better protect U.S. national security and help U.S. businesses thrive,” said McGuire of the Council on Foreign Relations. “But that type of partnership requires trust, which has to be earned.”