Great American AI Act Draft Sets $500 M Frontier Threshold and Federal Preemption

Why It Matters

The Great American AI Act draft could become the cornerstone of U.S. AI regulation, establishing the first federal revenue‑based threshold that separates large frontier developers from smaller innovators. By preempting state development rules for three years, the bill aims to create a uniform compliance landscape, reducing legal uncertainty for multinational AI firms while preserving state authority over deployment risks. The creation of CAISI and its audit framework introduces a new compliance market, potentially driving investment in AI governance services and shaping how companies document model safety. If enacted, the legislation would influence global AI policy, as foreign firms seeking to operate in the United States would need to align with the same standards. The $500 million threshold also raises questions about how revenue is measured and reported, potentially prompting firms to restructure or segment business units to stay below the regulatory line. The balance between fostering innovation and ensuring safety will be a key test for policymakers.

Key Takeaways

• Discussion draft released June 4, 2026 by Reps. Obernolte and Trahan
• Defines “large frontier developer” as any AI firm with >$500 M annual revenue
• Imposes a three‑year federal preemption of state AI development laws
• Creates Center for AI Standards and Innovation with $100 M annual funding (2027‑2029)
• Targets major AI firms—Anthropic, OpenAI, Google DeepMind, xAI, Meta—while exempting startups

Pulse Analysis

The Great American AI Act draft marks a decisive shift from the fragmented state‑level approach that has characterized U.S. AI policy to a more centralized, revenue‑driven framework. By anchoring obligations to a $500 million revenue threshold, Congress is effectively drawing a line between the industry’s “big‑ticket” players and the broader ecosystem of innovators. This mirrors the regulatory logic used in financial services, where systemic risk is tied to size, and suggests policymakers view frontier AI models as a potential source of systemic societal risk.

The three‑year preemption clause is a strategic gamble. It gives the federal government breathing room to develop a comprehensive rulebook while temporarily shielding developers from a cascade of state regulations that could vary wildly in scope and enforcement. However, the sunset provision also acknowledges that a permanent preemption would be politically untenable. Stakeholders will likely lobby intensely during this window, shaping the eventual federal standards and possibly influencing the composition of CAISI’s oversight board.

From a market perspective, the draft could catalyze a new compliance industry. Companies will need to invest in audit readiness, risk‑assessment tooling, and legal counsel to navigate both the federal obligations and the surviving state deployment laws. Smaller firms may see an advantage, as the exemption could allow them to innovate without the overhead of federal certification, potentially accelerating open‑source and niche AI solutions. Ultimately, the GAAIA’s trajectory will hinge on how Congress balances the desire for uniformity with the need to preserve state-level consumer protections, a tension that will define U.S. AI governance for years to come.