AI Labs Break Lobbying Records, With Anthropic Ahead of Nvidia

The AI labs are learning to spend money in Washington, and the second-quarter filings show just how quickly. Anthropic reported $1.97 million in federal lobbying between April and June, a 26 percent jump from the previous quarter and enough to pull it ahead of Nvidia, the chipmaker whose hardware underpins the entire industry. OpenAI spent $1.2 million over the same three months, up nearly 18 percent. Together the two labs put a record $3.17 million into influencing federal policy in a single quarter.

The timing is not hard to read. Anthropic's spending climbed sharply after the Commerce Department pulled two of its flagship models offline in June under an export-control order, the first time a frontier model had been switched off by regulators. A company that has just watched Washington reach into its product line has every reason to invest in being heard there. Federal filings show the labs concentrating on cybersecurity, copyright, cloud computing, and defense contracts.

Still small money, for now

It is worth keeping the numbers in proportion. Anthropic and OpenAI are outspending each other at a record pace, but they are still a rounding error next to the established giants. Meta remained the largest spender in the group at $5.99 million, with Amazon steady at $4.36 million and Alphabet up 25 percent to $3.57 million. The AI-native labs are climbing fast from a low base, not dominating the field.

What the trend tells you is where these companies now believe their real risks sit. A year ago the fights that mattered were about benchmarks and funding rounds. The bills came due in Washington instead, in the form of export rules, safety-testing regimes, and the growing patchwork of state AI laws. The labs raised billions to build the models. They are now paying, in the millions, to shape the rules those models have to live under. It is a familiar arc for any industry that suddenly matters, and it usually only goes one way.