What is the GAIN AI Act?

The GAIN AI Act is a proposed amendment to the U.S. national defense authorization bill that targets advanced AI chips produced by U.S. firms such as Nvidia and AMD. Under the bill:

• U.S. buyers would receive a 15-day right of first refusal (ROFR) on high-performance AI chips before those chips can be exported to specified foreign countries (notably adversaries, e.g., China).
• The intention is to prioritize domestic demand (i.e., U.S. firms and users) and thereby safeguard American access to the technology while placing constraints on overseas sales.
• The bill is pitched as both an industrial-policy move and a national-security measure – seeking to protect the U.S.’s technological edge in AI hardware.

In short: before a U.S. chipmaker exports a qualifying AI chip overseas under the GAIN AI Act, it must offer U.S. domestic buyers the chance to buy first, certify that there are no outstanding domestic backorders, and provide data on pricing/customer terms.

Why does it matter?

• The U.S. government is already employing export-controls on advanced semiconductors and AI hardware to restrict adversary access and maintain domestic leadership in AI and compute.
• Supporters of GAIN argue that if U.S. firms are selling large volumes of advanced AI chips abroad, then domestic demand might be underserved, and adverse actors might gain compute power that e.g. China could use for both commercial and strategic ends.
• So the bill aims to give U.S. buyers priority, and restrict exports – or at least condition them – when domestic supply might be constrained.

What are the problems with the GAIN AI Act?

It conflates national security with industrial policy

While framed as a security measure, many analysts argue the bill is really about “America First” industrial policy: favoring U.S. buyers and restricting exports for strategic economic gain. If a bill is truly grounded in national-security risk, you’d expect blanket bans on exports to high-risk countries rather than creating complex rules about ROFR and domestic priority.

It may harm the very innovation and competitiveness it seeks to protect

• Because the semiconductor/AI‐hardware business relies on global scale (selling internationally helps amortize R&D and fixed costs), restricting exports or forcing prioritization can undermine the economics of U.S. firms.
• Some industry voices say the premise  –  that U.S. domestic buyers are being deprived or diverted in favor of exports  –  is flawed. For example, Nvidia states that it “never deprives American customers in order to serve the rest of the world” and that the bill “addresses a problem that does not exist.”
• Over time, forcing U.S. firms into constrained global markets may reduce investment, slow innovation, increase unit cost, and give competitors (especially China) the chance to gain ground.

It introduces complexity, regulatory risk and unintended consequences

• The act would require disclosure of confidential pricing, customer data and terms of sale – essentially sharing competitive information with government or third-parties, which could lead to downstream collusion, reduced bargaining power, or interference with commercial strategy.
• U.S. exporters might face increased red tape, longer approval processes, and diminished global competitiveness. Meanwhile, adversaries might gain intelligence about the U.S. industry’s pricing, backlog, and supply chain.
• The bill may restrict supply abroad, raising prices for foreign buyers, but reduce the scale and revenue for U.S. chipmakers  –  again undermining the U.S. industry.
• Moreover, export controls targeting specific hardware are increasingly vulnerable to substitution, indigenization, and workaround by adversary nations (for example China developing its own supply chain). Export restrictions alone may lose effectiveness over time.

It risks locking in a protectionist framework

Once enacted, these rules may become hard to unwind (as past examples like the Jones Act in maritime law illustrate). A short-term policy may become extended into long-term industrial policy, with negative effects for U.S. competitiveness.

Final thoughts

While the impulse behind the GAIN AI Act – ensuring U.S. access to cutting-edge AI hardware and limiting tech flows to adversaries – is understandable, the mechanism chosen is problematic. By prioritizing domestic buyers via ROFR, mandating disclosures, and constraining exports in a globally-interconnected supply chain, the bill runs the risk of undermining both innovation and national security in the medium term.

Rather than a buyer-first right and blanket export constraints, policy might better focus on: improving U.S. manufacturing capacity, easing domestic regulatory bottlenecks (e.g., for data centers, energy supply, infrastructure), and using smarter, targeted export controls rather than broad domestic-priority mandates. The U.S. must compete globally, not just restrict. Otherwise, we risk giving the lead to someone else – even when we think we’re protecting ourselves.