HR 9925
To provide for Federal oversight of the development and deployment of frontier artificial intelligence in interstate and foreign commerce, and for other purposes.
TL;DR
Rep. Jay Obernolte (R-CA) introduced HR 9925 to create federal oversight of 'frontier' AI models, the largest and most powerful systems being built by companies like OpenAI, Anthropic, and Google. The bill would establish federal rules for how these advanced AI systems are developed and deployed across state lines and internationally, likely including safety testing and reporting requirements before deployment.
How This Might Impact Your Business
Frontier AI developers (companies training the largest models, think OpenAI, Anthropic, Google DeepMind, Meta, xAI) would face new federal registration, safety testing, and deployment rules.
Enterprises deploying these models through APIs (in healthcare, finance, legal tech, customer service) could inherit compliance obligations tied to their vendors' certifications.
Small and mid-size AI startups using open-source or smaller models likely fall outside the scope, since 'frontier' typically means models above a compute or capability threshold.
Interstate and foreign commerce framing means cloud providers (AWS, Azure, GCP) hosting frontier models would be pulled into the compliance chain.
No specific penalties, thresholds, or timelines are public yet; the bill sits in the Energy and Commerce and Science, Space, and Technology committees, meaning details will be shaped in markup.
Bipartisan interest in frontier AI oversight (Obernolte co-chairs the House AI Task Force) makes some version of this framework likely to advance, even if this specific bill stalls.
Government contractors and federal agencies procuring AI could see new certification requirements flow down into vendor agreements.
What Should You Do
Ask your AI vendors (OpenAI, Anthropic, Google, Microsoft) how they are tracking federal frontier AI legislation and what their compliance posture would be.
Inventory which of your internal AI use cases rely on frontier models versus smaller or open-source models, since regulatory exposure will differ sharply.
Assign someone in legal or policy to monitor Energy and Commerce and Science, Space, and Technology committee activity on HR 9925 through the end of this Congress.
Review vendor contracts for language on regulatory change, safety testing disclosures, and the right to switch models if compliance costs get passed through.
If you sell AI-powered products to federal agencies, brief your government affairs team now on likely certification and reporting flow-downs.
Who It Affects
Sponsors
Status Timeline
committee
Referred to the Committee on Energy and Commerce, and in addition to the Committee on Science, Space, and Technology, for a period to be subsequently determined by the Speaker, in each case for consideration of such provisions as fall within the jurisdiction of the committee concerned.
July 23, 2026