Federal lobbying disclosures published this week show that Anthropic spent $1.6 million pressing its case on Capitol Hill in the first quarter of 2026, eclipsing OpenAI's $1 million over the same period. Both figures are record highs for the two AI labs, according to reporting from Axios on Tuesday, and underline how quickly policy has become a front-line expense for frontier model developers.
A record quarter on both sides
Anthropic's filing is not just a new personal best — it also flips the longtime assumption that OpenAI, with its much larger historical footprint in Washington, would dominate federal spending. OpenAI has crossed $25 billion in annualized revenue, while Anthropic reported topping $30 billion in run-rate revenue earlier this month, yet it is Anthropic — long the smaller government-affairs operation of the two — that opened the widest checkbook this quarter. OpenAI's $1 million tally is likewise reportedly the most it has ever reported in a single quarter, a sign that every major lab is now scaling up its government-affairs footprint rather than pulling back.
What each company is lobbying on
The disclosures offer an unusually direct look at each company's policy priorities. Anthropic's lobbyists reported contacts on AI procurement, Defense Department procurement, supply chain risk and acceptable use policy — the exact terrain of its ongoing dispute with the Pentagon over how Claude can be deployed in classified and operational settings. The filing also lists AI and national security, export controls, legislation, energy infrastructure, supply chain and permitting.
OpenAI's Q1 2026 report is narrower. It cites AI and copyright, cybersecurity, cloud computing and infrastructure. Copyright has been a live issue for the company as publisher lawsuits grind through federal courts, and infrastructure and cloud computing language tracks with the company's push for favorable siting and permitting rules around its Stargate-scale data-center build-out.
Policy as a core product line
The lobbying surge arrives alongside a parallel escalation in political spending. Anthropic has backed the Public First Action bipartisan 501(c)(4) nonprofit with a $20 million commitment, while OpenAI-aligned donors — including co-founder Greg Brockman and Andreessen Horowitz partners — have seeded a rival vehicle, "Leading the Future," which had already raised $125 million as of January 2026 for the 2026 midterms. Direct federal lobbying and outside political spending are tracked separately, but both lines are now moving in the same direction: up.
Implications for the AI policy map
For lawmakers, the Q1 disclosures confirm that the policy debate is no longer theoretical. Pentagon acceptable-use rules, state-level liability bills, copyright carve-outs and data-center permitting are all live fights, and the companies with the most exposure are staffing up accordingly. Expect the Q2 filings, due in July, to reset these records again — and to bring Google, Meta and xAI more clearly into the comparison as enterprise and defense deals pull every frontier lab deeper into the legislative process.
