Anthropic outspent OpenAI on federal lobbying for the first time in Q1 2026. Meta led all companies in 2025 with $26.29 million in lobbying — the most of any company in any industry. More than 3,570 federal lobbyists worked AI issues in 2025, equal to roughly one in four federal lobbyists in the United States.
EPR Research and 5W today released The AI Lobbying Industry Map 2026, the first systematic mapping of how AI companies and their trade associations are deploying federal lobbying capacity. The Study draws from the Senate Office of Public Records, the House Lobbying Disclosure database, OpenSecrets, Issue One, and Public Citizen's analysis of LDA filings.
The Top-Line Numbers
Top tech and AI companies spent more than $100 million on federal lobbying in 2025 — the first time the category exceeded that figure. Per Bloomberg's compilation of federal LDA filings:
Company2025 Federal Lobbying SpendMeta$26.29M (highest of any company, any industry)Amazon$17.89MAlphabet (Google)$13.10MMicrosoft$9.36MNvidia$4.9M (7x its 2024 spend)OpenAI~$3M (up from $1.76M in 2024 and $260K in 2023)Anthropic$720K in 2024, accelerating into 2026
The Q1 2026 Inflection Point
Per Issue One's review of newly filed federal lobbying reports, 11 key technology, social media, and AI companies and their associated trade associations together spent a combined $20 million on federal lobbying between January and March 2026 — an average of about $226,000 per day.
The most striking shift: Anthropic outspent OpenAI on federal lobbying for the first time ever in Q1 2026, reporting nearly $1.6 million in spend — a 333 percent increase from the $360,000 it spent in Q1 2025, an average of roughly $17,000 per day. OpenAI reported $1 million for the quarter, an 82 percent year-over-year increase. Nvidia reported $1.3 million in Q1 2026 federal lobbying, up 37 percent year-over-year. Snap reported $400,000 and X reported $190,000 — both up modestly. ByteDance, parent of TikTok, went the other direction: $1.6 million in Q1 2026, down roughly 45 percent from $2.8 million in Q1 2025, following the January 2026 deal that restructured U.S. ownership through Oracle, Silver Lake, and MGX.
The Lobbyist Headcount
According to Issue One, the six largest tech, social media, and AI companies — Alphabet, Anthropic, Meta, Microsoft, Nvidia, and OpenAI — collectively hired 307 lobbyists during Q1 2026. That is the equivalent of slightly more than one lobbyist for every two members of Congress.
Alphabet alone hired 88 lobbyists — roughly one for every six members of Congress.
Meta hired 86 lobbyists — also roughly one for every six members.
30 of the 307 lobbyists work for private K Street firms representing two or more of these six companies.
Per Public Citizen's analysis of House lobbying disclosure data, 3,570 federal lobbyists reported lobbying on AI issues at least once in 2025 — more than 25 percent of the entire federal lobbying corps. Three years ago, that figure was 11 percent. The number of unique lobbyist–client relationships on AI issues has grown 265 percent over the past three years. The number of data center lobbyists alone has grown nearly 500 percent.
If AI were classified as a single OpenSecrets issue area, it would rank fourth in the United States behind only the federal budget and appropriations, taxes, and health.
Strategic Pivots Visible in the Filings
Anthropic has pivoted into healthcare AI and Defense Department procurement. Q1 2026 disclosures reference the Healthcare AI Accountability Act (S. 4178) and DoD acquisition policy. Anthropic retained Ballard Partners to pursue Pentagon AI procurement contracts following public disputes over Anthropic's acceptable use policy and classified-setting deployment.
OpenAI has pivoted into energy and infrastructure, focused on the Stargate joint venture with SoftBank and Oracle and on data center permitting reform, including the Data Center Energy Efficiency Act (H.R. 4425). OpenAI's lobbying spend has grown roughly 7x year-over-year on this push.
Meta has the most aggressive state-level operation. Per CalMatters, Meta spent at least $4.6 million lobbying California state officials in 2025, more than in any prior year since it began Sacramento advocacy in 2010, and transferred $20 million to a new political committee created to support candidates favoring fewer AI regulations. Meta also made 34 behested payments totaling $1,245,000 at the request of California officials in 2025.
The Revolving Door
OpenAI hired Reginald Babin, former legal counsel to then–Senate Majority Leader Chuck Schumer, as a registered lobbyist. OpenAI later hired veteran political operative Chris Lehane as VP of policy and global head of policy. Meta recruited a former Trump adviser as president and vice chairman in 2025. OpenAI hired a former Trump adviser to lead its global energy policy. OpenAI President Greg Brockman and his wife together gave $25 million to a Trump-aligned political action committee.
Anthropic brought on its first in-house lobbyist, Department of Justice alum Rachel Appleton, in late 2024. The Data Center Coalition — which counts OpenAI and Anthropic among its members — reported $420,000 in Q1 2026 lobbying, more than tripling from $123,000 in Q1 2025.
Methodology
The Study draws from Senate Office of Public Records LDA filings (LD-1 registrations and LD-2 quarterly reports), House lobbying disclosure data, OpenSecrets aggregations, Issue One filings reviews, Public Citizen analysis of AI-issue lobbying volume, Bloomberg compilations of company-level totals, and CalMatters analysis of California state lobbying. Coverage spans federal lobbying for FY2023 through Q1 2026, plus California state lobbying for 2024 and 2025.
Disclosed Limits
Lobbying disclosure undercounts: not all advocacy meets the LDA registration threshold, dark-money advocacy is excluded entirely, and the February 2025 DOJ memorandum from Attorney General Pam Bondi limiting FARA enforcement may suppress future filings. The Q1 2026 figures are based on initial filings and may be revised.
Why It Matters
For 25 years, the most consequential algorithm in communications was Google PageRank. The most consequential algorithm in U.S. federal policymaking is now whichever AI model becomes the dominant inference engine inside government. The companies fighting to set those rules are deploying lobbyists at the highest rate in the history of federal disclosure. Every public affairs firm, every comms team at a regulated company, every reporter covering AI policy, and every state attorney general's office has reason to track this map.
