Super PAC networks affiliated with Anthropic and OpenAI have spent more than $37 million on 2026 congressional primaries to influence federal AI policy. According to Federal Election Commission (FEC) filings, OpenAI-linked groups Think Big PAC and American Mission have spent $14.1 million and $8 million respectively through June 15. Anthropic-aligned groups, Jobs and Democracy and Defending Our Values, have spent $11 million and $5.2 million during the same period.

The organizations primarily support candidates from both major parties but have conflicted in the Manhattan Democratic primary race to replace Rep. Jerry Nadler. The contest centers on New York Assemblymember Alex Bores, who authored the RAISE Act, a law signed in December 2025 requiring large AI developers to report safety protocols and incidents to state regulators within 72 hours. The OpenAI-affiliated Leading the Future network has spent over $7 million attacking Bores, while Anthropic-backed groups have provided roughly equivalent funds for his defense.

The spending reflects a fundamental policy disagreement between the two companies. OpenAI advocates for AI regulation exclusively at the federal level, while Anthropic supports state-level mandates similar to those enacted in New York and California. This influence extends beyond urban races; in Montana's 1st Congressional District Republican primary held on June 2, an OpenAI-linked super PAC spent approximately $877,000 to boost candidate Aaron Flint, who ultimately defeated former state Sen. Al Olszewski.

These political activities coincide with increasing federal scrutiny of AI companies. On June 12, Commerce Secretary Howard Lutnick mandated that Anthropic disable foreign access to its Mythos 5 and Fable 5 models under new export control regulations. This move aligns with a broader Trump administration executive order from December 2025, which directs federal agencies to challenge state-level AI regulations that conflict with national policy. Despite the significant investment, both networks have largely avoided mentioning specific AI technologies in their campaign advertisements, preferring to focus on candidate biographies and broader policy themes.

Super PAC networks affiliated with Anthropic and OpenAI have spent more than $37 million on 2026 congressional primaries to influence federal AI policy. According to Federal Election Commission (FEC) filings, OpenAI-linked groups Think Big PAC and American Mission have spent $14.1 million and $8 million respectively through June 15. Anthropic-aligned groups, Jobs and Democracy and Defending Our Values, have spent $11 million and $5.2 million during the same period.

The organizations primarily support candidates from both major parties but have conflicted in the Manhattan Democratic primary race to replace Rep. Jerry Nadler. The contest centers on New York Assemblymember Alex Bores, who authored the RAISE Act, a law signed in December 2025 requiring large AI developers to report safety protocols and incidents to state regulators within 72 hours. The OpenAI-affiliated Leading the Future network has spent over $7 million attacking Bores, while Anthropic-backed groups have provided roughly equivalent funds for his defense.

The spending reflects a fundamental policy disagreement between the two companies. OpenAI advocates for AI regulation exclusively at the federal level, while Anthropic supports state-level mandates similar to those enacted in New York and California. This influence extends beyond urban races; in Montana's 1st Congressional District Republican primary held on June 2, an OpenAI-linked super PAC spent approximately $877,000 to boost candidate Aaron Flint, who ultimately defeated former state Sen. Al Olszewski.

These political activities coincide with increasing federal scrutiny of AI companies. On June 12, Commerce Secretary Howard Lutnick mandated that Anthropic disable foreign access to its Mythos 5 and Fable 5 models under new export control regulations. This move aligns with a broader Trump administration executive order from December 2025, which directs federal agencies to challenge state-level AI regulations that conflict with national policy. Despite the significant investment, both networks have largely avoided mentioning specific AI technologies in their campaign advertisements, preferring to focus on candidate biographies and broader policy themes.