Antitrust abuse threatens America’s AI dominance

On July 14, 2025, Jensen Huang, CEO of Santa Clara-based NVIDIA, and President Donald Trump agreed to allow NVIDIA to resume its exports to China amidst ongoing US-China trade disputes.

By the next day, the stock had surged 4 percent in response to the news. The resultant market cap of $4.4 trillion made NVIDIA the highest market cap company in the world–and the first over $4 trillion–with an estimated 85 percent of the global AI chip market.

NVIDIA’s competitive edge lies in its superior graphics processing units (GPUs), the essential hardware infrastructure that enables AI machine learning and processing. NVIDIA’s emphasis on gaming and 3D rendering made its chips almost accidentally perfect for AI applications and in turn, vital to propelling the US in what some regard as a global AI race. Ill-advised export restrictions on NVIDIA’s H20 GPU initially lost the company $5.5 billion in stranded inventory. Now, though, NVIDIA plans to export the GPUs to China in appropriately downgraded form (given national security concerns) to expand its market base once again.

The company’s latest economic surge is a continuation of the momentum it has built since ChatGPT’s launch in 2022.

NVIDIA’s 85 percent market share still remains vulnerable to ongoing antitrust adventurism, which is perhaps even more ill-advised than disruptive trade policy. The possession of large market share does not automatically place a firm in violation of antitrust laws, but policymakers always seem to have difficulty with that realization. Regulators should be particularly wary of overzealously enforcing antitrust law against NVIDIA, which can harm US competitiveness abroad as surely as can trade restrictions.

Under the Biden administration in 2024, the US Department of Justice (DOJ) launched an antitrust probe into NVIDIA to discern whether the company was charging higher prices to customers who buy AI chips from competitors. NVIDIA has also received scrutiny from international antitrust authorities. In 2003, the European Commission investigated the $700 million acquisition of Israeli AI startup Run:ai, but later approved the deal. Even China launched an antitrust probe into the 2019 acquisition of Mellanox for $7 billion.

Large technology companies continue to face unwarranted antitrust pressure from both Republicans and Democrats, but the stated reasons for that scrutiny differ–from alleged abuse of advertising networks and app store restrictions, to self-preferencing in search and cloud partnerships. Yet even those relationships and sectors remain vibrant demonstrations of competitive free enterprise rather than monopoly. And while Microsoft, Apple, Google, and Meta rely on NVIDIA’s GPU chips at various levels of AI development, those companies themselves are pouring billions of dollars into research and development, including into custom chip design.  

Paradoxically, the current administration is maintaining some antitrust lawsuits brought against tech companies during the Biden administration. Needless to say, policymakers will find it challenging to fulfil their aims of leading the AI race when the American companies responsible for the technology’s development continue to face down antitrust lawsuits brought forward by the Federal Trade Commission (FTC) and the DOJ. That is hardly incentivizing AI growth.

Better steps include deregulation as well as freer trade. In January 2025, President Trump issued EO 14179 (Removing Barriers to American Leadership in AI) which focuses on deregulating and fostering AI development.

More recent decrees from the White House include the release of America’s AI Action plan and a fresh series of executive orders such as Preventing Woke AI in the Federal government, Accelerating Federal Permitting of Data Center Infrastructure and Promoting Export of the American AI Technology Stack. While these provisions are a mixed bag and do contain some discordant regulatory ingredients, they also do include important new liberalizations such as the fast-tracking of permitting for AI data centers, prevention of overzealous state interference with AI, a leash on the FTC, and limitations on AI hardware export restrictions.

While the Trump administration’s new executive actions aim at relieving the AI sector of certain barriers, this alone will not suffice. Congress still must act to lock in permanent deregulation, not just for AI but for the economy as a whole. But for this sector to truly breathe and expand, the United States absolutely must abandon its constant deployment of antitrust regulation against leading tech innovators. This starts with recognizing NVIDIA as a success story.

AI development requires extraordinary capital investments and if the US acts now to free them up and create certainty, the AI race can be won.