Three Big Questions Washington Must Answer to Secure America's AI Future

Policymakers in Washington are confronting three interlinked challenges as artificial intelligence advances rapidly: how to set federal rules that work across 50 states, how to compete with China for AI leadership and standards, and how to meet the vast energy and infrastructure needs of AI data centers. Current and former lawmakers, industry leaders and experts debated these issues at a recent "AI in America Summit."

1. State preemption and the fight over regulation

Lawmakers remain sharply divided over whether Congress should preempt state AI laws with a uniform federal standard. Several states — including California, Texas, Florida, Maryland and Colorado — have already passed statutes addressing different AI risks. Critics warn that a patchwork of state rules could stifle innovation, while others say federal preemption must include substantive protections.

Sen. Mike Rounds (R-S.D.): "The fifty states are laboratories showing Congress where people are concerned about how we regulate artificial intelligence."

A proposed 10-year moratorium on state-level AI governance was added to a GOP tax and spending bill in May but the Senate voted 99-1 in July to remove it. House leaders later tried to reinsert the moratorium into the National Defense Authorization Act, an effort that faced opposition inside the GOP caucus.

Former Rep. Garret Graves (R-La.), co-chair of the AI Infrastructure Coalition, argued that heavy-handed or duplicative regulation could "undermine the entire AI industry," and suggested federal agencies such as the Federal Communications Commission and the Environmental Protection Agency could apply existing tools to oversee aspects of AI. Rep. Ted Lieu (D-Calif.) said he would accept federal preemption only if it established a strong, reasonable standard rather than an empty one.

2. The U.S.–China race for AI leadership

Lawmakers said the United States must shape the technical standards and norms that govern AI development. Many fear that if China defines those rails, AI systems worldwide could reflect different priorities and values.

Sen. Mike Rounds: "Who do you want to determine the standards and the norms long term for the development of AI? For most of us, the answer is that it has to be the U.S. and its allies."

Concerns intensified after the Chinese startup DeepSeek released its R1 model earlier this year. The company claimed the model could rival leading American systems at a fraction of the cost, and its app briefly surged to the top of app stores. Observers called the moment an "AI Sputnik," though DeepSeek's model also demonstrated state-aligned content limits — declining to respond on politically sensitive topics such as Tiananmen Square — underscoring how different political systems can shape AI behavior.

Participants warned that preserving American values in AI development requires sustained investment and strategic policy choices.

3. Energy, infrastructure and the practical limits of scaling AI

Many panelists said energy and land are immediate, tangible constraints on U.S. competitiveness. Large-scale AI training and inference require dense clusters of GPUs inside data centers that consume substantial electricity and water.

As of May 2021, the United States had more than 2,600 data centers. A 2021 study found roughly 20% of U.S. data centers were located in watersheds under moderate to high stress. Rising electricity prices — reported at 5.1% year-over-year in one recent September reading — and projections from the Open Energy Outlook that data-center growth could increase national electricity costs by about 8% by 2030 add urgency to the issue. The Energy Department has warned about a rising risk of blackouts amid rapid data center expansion and efforts to reshore manufacturing.

Fred Thiel, CEO of MARA: "The biggest challenge to AI today is land and power. If you don’t have energy, you can’t have AI data centers. Thousands of GPUs sit idle waiting for utilities to flip the switch."

Speakers called for coordinated policy on siting, grid upgrades, permitting and incentives for low-cost, low-carbon energy to support responsible AI growth.

What comes next

Lawmakers and industry leaders left the summit agreeing that Congress and regulators face a narrow window to craft policy that balances innovation, national security and public protections. Key near-term choices — whether to preempt state laws, how to respond to international competitors, and how to scale energy and infrastructure — will shape the U.S. AI landscape for years to come.