Federal lobbying disclosures reviewed by Semafor show that OpenAI and Anthropic met with lawmakers in the third quarter to discuss how AI companies might share information with one another without violating antitrust rules. The filings indicate both firms engaged on issues of security risk coordination and information-sharing protections in Washington.
Specific topics listed in filings
Anthropic’s lobbying disclosure specifically names the Collaboration on Adversarial Threats and Security Risks Act, a bill that would permit labs to coordinate with each other on security risks. The filing notes the legislation could be folded into must-pass defense policy legislation.
OpenAI’s lobbying form is less detailed, listing “AI-security collaboration and information-sharing, including proposed antitrust protections” among topics discussed. Lobbying forms do not indicate whether the companies supported or opposed the proposals discussed during meetings.
Context: antitrust debate and calls to slow development
Antitrust lobbying has gained attention against the backdrop of recent industry discussions about a coordinated slowdown in AI development. That debate centers on whether companies should be allowed a legal carve-out to coordinate on model pacing, or whether such coordination can and should happen voluntarily.
Chris Lehane, OpenAI’s head of global affairs, has argued that an antitrust exemption specifically for model pacing is not necessary. By contrast, officials in the Trump administration criticized AI leaders for proposing such a carve-out, saying companies are capable of deciding among themselves to slow down without government approval.
Lobbying spending: Meta leads, Nvidia posts a record quarter
Senate lobbying filings show Meta spent $5.9 million on lobbying in the third quarter of 2026, the largest amount among big tech firms in that period. Observers note the AI race has intensified and high-profile security incidents have added pressure to the AI-safety debate in Washington.
Google, Amazon, and Apple spent less than Meta but more than many AI and chip firms. Nvidia reported $2.2 million in lobbying for the quarter, its highest-ever quarterly total, reflecting the chipmaker’s continued dominance and influence within the Trump administration.
Rising spending by AI startups
Semafor’s review also found that OpenAI and Anthropic have increased their lobbying expenditures sharply this year: OpenAI’s spending nearly doubled and Anthropic’s nearly tripled in the first half of the year. Both companies are engaging lawmakers on a range of policy topics, including proposed AI regulations, export controls, and data center policy.
Summary
Disclosures reviewed by Semafor show OpenAI and Anthropic discussed antitrust-safe information-sharing and coordination on security risks with lawmakers in Q3, while Meta was the largest lobby spender in Q3 2026 at $5.9 million. The filings underscore how antitrust, security, and lobbying are intersecting as regulators and companies debate how to manage AI development and safety.
