
Anthropic Just Outspent OpenAI on Washington Lobbyists. A Year Ago It Was Spending Almost Nothing.
Anthropic dropped $1.6M on lobbying in Q1. OpenAI spent $1.24M. A year ago, Anthropic spent $360K. What changed: the Pentagon banned them.
The AI Post newsroom — delivering AI news at the speed of intelligence.
Anthropic, the AI company Washington loves to hate, just became the biggest AI lobbying spender in the United States.
The company spent $1.6 million on federal lobbying in the first quarter of 2026, according to new disclosures first reported by Axios. That is a 344% jump from the $360,000 it spent in Q1 2025. It is also more than what OpenAI spent in the same quarter ($1.24 million), making Q1 2026 the first time Anthropic has outspent its larger rival in Washington.
This is a straight line from our coverage going back six weeks. The Pentagon designated Anthropic a supply chain risk in February. Trump told reporters in early April he had "no idea" who they were. Then Dario Amodei walked into the White House on April 17 and had a "productive" meeting with Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent. That meeting did not happen by accident. It happened because Anthropic hired Ballard Partners, the lobbying firm where Wiles worked for years, specifically for advocacy on Department of War procurement.
$1.6 million is what buys you a chair at that table.
Meta Still Wins Dollar For Dollar
For perspective: Meta spent $7.6 million on lobbying in Q1. Google parent Alphabet spent $2.9 million. Microsoft, Amazon, Apple all cleared $2 million. The AI labs are still punching below Big Tech weight, but the gap is closing fast.
What makes Anthropic's number remarkable is the velocity. Most lobbying budgets grow 10 to 20 percent year on year. A 344 percent jump in one quarter is a crisis response. It is a company that woke up one morning and realized Washington could wipe out a material share of its future revenue with a signature on a memo.
And that is roughly what happened.
The Pentagon Ban Changed The Math
Before February, Anthropic could afford to play the "AI safety is our brand" game. The pitch to regulators was that Anthropic was the responsible lab, the one that worked with governments rather than against them. That pitch relied on the assumption that working with the federal government was not, itself, a regulatory risk.
Then the Pentagon cut them off. Defense Department contractors were told to remove Mythos Preview from their workflows. The supply chain risk designation, once applied, becomes contagious: banks, insurers, and federal contractors all have procurement rules that freeze vendors with that label. Anthropic was staring at revenue lockout from the single biggest enterprise buyer of AI in the country.
$1.6 million looks cheap next to that outcome.
OpenAI's Number Is The Real Story
Here is what is easy to miss. OpenAI spent $1.24 million. That is up, but not dramatically. OpenAI has been running a steady Washington operation for two years. Sam Altman testified to Congress in 2023 and has been in lobbying mode ever since. Its spending is sustained, not reactive.
Anthropic's spending is reactive. The 344% jump is the signature of a company that just had its business model threatened and is now trying to buy its way back to neutral. The question is whether it worked. The Wiles meeting, the thaw with Bessent, the rumored reinstatement of Pentagon procurement, all point to yes. Amodei got the meeting. He got the room. He is still not getting Pentagon contracts, but the door is open.
What We Are Watching Next
Three things.
First, whether Ballard Partners delivers the Pentagon reversal. If Anthropic's supply chain risk designation gets lifted within the next 60 days, the $1.6 million was the best money the company has ever spent. If not, the spending ratchets up again in Q2.
Second, whether OpenAI responds. OpenAI has always viewed Washington as a steady-state relationship, not a transactional one. If Anthropic's aggressive spending starts winning specific carveouts (Mythos Preview access for certain agencies, a federal AI procurement framework favorable to smaller labs), OpenAI will have to match.
Third, the IPO optics. Both companies are heading to public markets within the next 18 months. Federal lobbying disclosures are required reading for institutional investors. A company that spends 344% more on Washington in a single quarter looks like a company with regulatory exposure. That is a story IPO underwriters will have to answer.
Anthropic just told Wall Street where the risk lives. It is spending $1.6 million a quarter trying to make that risk go away.
Source: Axios (primary, April 21 2026), via federal lobbying disclosure filings. Further context: CNBC, CNN, TechCrunch, TheNextWeb coverage of the April 17 Amodei/Wiles meeting.