Fortune Commentary: Nvidia, Export Controls, and the Growing Risk of AI Corruption
Key Takeaways
- •A 2025 panel convened by the UN Office on Drugs and Crime and the Organization of American States feared AI would enable corruption, but by November 2026 the corruption emerged in AI governance led by officials and executives rather than in the technology itself.
- •President Trump permitted Nvidia to sell the H200 chip to China on condition that the U.S. government receive a share of the sales, although the Congressional Research Service reported that the Constitution prohibits federal export taxes and that law bars BIS from collecting license fees.
- •Nvidia's then-General Counsel Tim Teter appeared to interfere in an investigation of the company's sales to Chinese and Malaysian entities, and senior officials including Commerce Secretary Howard Lutnick were warned about the risks but took no action.
- •Export controls are especially vulnerable to corruption because they concentrate discretionary licensing power, create artificial scarcity and high profits, and rely on opaque national security exemptions — vulnerabilities compounded by the Trump Administration's 2025 dismissal of many inspectors general.
- •The author argues export controls are rarely effective and calls on Congress to coordinate technology-transfer management with Western allies, even though Politico reports lawmakers face near-impossible political odds in supporting looser controls.

In 2025, the UN Office on Drugs and Crime and the Organization of American States co-convened a panel (slideshow & video) examining whether artificial intelligence is the biggest corruption risk in history. By November 2026, the panelists' deepest fears about the AI-corruption relationship had come true — though not in the way they expected.
The panelists had feared that AI would enable corruption by making it easier to launder money, pay bribes, or produce disinformation. Instead, they learned that the corruption of AI did not begin inside the technology itself, but with its leaders — heads of state, government officials, and tech CEOs — steering the direction of AI governance. America's approach to AI governance, which relies on export controls (government restrictions on the sale of sensitive technologies to foreign buyers) to regulate the supply of AI and its components, is fueling corruption today, using Transparency International's definition of the term: the abuse of entrusted authority for private gain.
In a Fortune commentary, Susan Ariel Aaronson argues that one company in particular — Nvidia — has become the poster child for corruption in the United States. Last year, the Trump Administration restricted sales of chips to China under U.S. law, out of fears that they would give the Chinese military an advantage or undermine U.S. leadership in AI. Nvidia's now-former CEO, Jensen Huang, used flattery and donations to influence the administration's AI export policies. In December, President Trump gave Nvidia permission to sell the H200, an older but still powerful AI chip, to China — as long as the U.S. government received a cut of the sales. Rival chipmaker AMD also received an exemption. In doing so, the U.S. gained a financial incentive for loosening controls on AI infrastructure exports to China. However, the Congressional Research Service reported that Article I, Section 9, Clause 5 of the U.S. Constitution "prohibits the federal imposition of export taxes and duties (not user fees)," and that 50 U.S.C. §4815(c) "prohibits BIS from collecting fees for considering or issuing export licenses." BIS refers to the Bureau of Industry and Security, the Commerce Department agency that administers U.S. export controls.
Additionally, last year Nvidia's General Counsel, Tim Teter, seemed to interfere in an investigation of Nvidia's sales to Chinese and Malaysian entities. And senior administration officials, including Commerce Secretary Howard Lutnick, were warned about the dangers surrounding Nvidia and its now-former CEO but took no action. This corruption and inaction, the author contends, puts U.S. national security and the American people at risk.
While export controls may be useful and necessary, they are particularly vulnerable to corruption because they sit at the intersection of high-value restricted goods, bureaucratic gatekeeping in customs, business, and trade policymaking agencies, and complex cross-border supply chains. When governments restrict access to lucrative technologies, including dual-use items (goods with both civilian and military applications) such as chips or AI, they can create massive financial incentives for bad actors to bribe officials or exploit regulatory blind spots to bypass the rules (OECD; Transparency International UK). Export control systems rely on government officials to issue licenses, classify goods, and clear shipments through customs, and this concentration of discretionary power allows corrupt officials to solicit bribes in exchange for faster approvals or looking the other way. At the same time, U.S. agencies struggle to coordinate trade, customs, and anti-corruption efforts, which are controlled by different agencies.
Export controls also create conditions of artificial scarcity and high profits, particularly for dual-use items such as AI systems; some buyers may be willing to pay extremely high prices for a limited supply of AI services. Perhaps unsurprisingly, competitors of Nvidia that produce high-speed chips, such as Google, as well as OpenAI, Amazon, Anthropic, and Microsoft — three of Nvidia's biggest clients — support legislation that would further limit Nvidia's ability to export chips to China.
Finally, export controls often involve opaque national security exemptions. Because export decisions frequently involve classified intelligence or sensitive defense deals, licensing and procurement processes often lack transparency and accountability, making it difficult for independent watchdogs to spot irregularities — a problem compounded in 2025, when the Trump Administration fired many inspectors general (the independent officials who investigate waste, fraud, and abuse inside federal agencies), including at the Department of Commerce, along with anti-corruption prosecutors.
Given these problems, the author argues that export controls are not the best tool for preventing U.S. adversaries from acquiring advanced AI systems or chips, because they present many opportunities for corruption. Yet Politico has reported that it is almost impossible for members of Congress to support loosening export controls. Studies have shown that such controls are rarely effective (CSIS; CNAS), and Congress should therefore weigh a different approach. Moreover, current export control implementation can harm U.S. economic and technological competitiveness. While there are no easy answers, the commentary concludes, the U.S. should work with its Western allies to manage technology transfers (the movement of advanced chips, software, and technical know-how across borders) that might threaten shared security interests. Whether Congress acts on that recommendation, despite the political difficulty Politico describes, is the open question to watch.
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune. This story was originally featured on Fortune.com.