AI Labs Outspend the Old Guard in Washington as Lobbying Hits Records

Fresh second-quarter disclosures show AI developers pouring unprecedented money into federal lobbying. OpenAI and Anthropic together reported roughly $3.2M for the April–June quarter, up about 23% from the start of the year, with Anthropic accounting for close to $2M of that. Across the first half of 2026, Anthropic’s federal lobbying tab climbed to about $3.5M — nearly triple its year-earlier total — while OpenAI’s roughly $2.2M was almost double. The two now spend more influencing Washington than chipmaker Nvidia, even as several legacy tech and defense contractors held their spending flat or pulled back.

The money is aimed at a live policy fight over who writes the rules for AI. The central question is whether Congress should impose a single national framework that overrides the growing patchwork of state laws, and the leading labs are not aligned: OpenAI has leaned toward federal preemption while Anthropic has been more guarded, emphasizing safety standards and, in places, state-level rules. Disclosed lobbying touched export controls, cybersecurity, copyright, cloud computing, and defense procurement, and the spending is bleeding into election machinery — including large AI-aligned super PACs and public-education vehicles ahead of the 2026 midterms.

The significance is that AI companies, several of them still young and unprofitable, are now buying influence at the scale of entrenched incumbents. That signals both that binding regulation is seen as imminent and that the industry intends to shape it — raising familiar concerns about firms writing the rules they will be governed by, at a moment when state and federal approaches are diverging sharply.