OpenAI, the company behind ChatGPT, captured about 90% of all spending by the U.S. House of Representatives on AI tools during the year ending March 31, 2026. House records show Congress spent roughly $100,580 on ChatGPT across 798 separate purchases, out of about $113,740 in total AI spending by House offices, committees, and institutional accounts, according to records analyzed by CNBC and reported by TechCrunch (TechCrunch).

Anthropic, the company behind a rival AI assistant called Claude, came in a distant second at $13,160 across 37 purchases. No other AI tool showed up in the records with any meaningful spending. The gap is wide enough that the real question is not which AI tool Congress prefers, but whether any competitor is in the conversation at all.

Democratic offices spent $54,165 on AI tools, about three times the $15,782 spent by Republican offices. The rest came from committees and institutional accounts rather than individual member offices. These numbers do not include AI used through free accounts or AI features built into other software that Congress already pays for, so the real amount of AI use on Capitol Hill is larger than these figures show (TechCrunch).

Congressional staff use ChatGPT and similar tools to write memos, summarize legislation, respond to citizens, prepare for hearings, research policy topics, and draft social media posts. In other words, AI has worked its way into nearly every kind of routine work that keeps a congressional office running.

The House of Representatives publishes its receipts and spending for representatives, committees, leadership, and officers, along with disclosures covering income, foreign travel, privately reimbursed travel, and legal expenses (house.gov). The AI spending came to light through this public reporting system, which lets anyone search individual transactions but does not group them by vendor or category.

Separately, OpenAI deployed 324 lobbyists during the second quarter of 2026, according to an analysis by Issue One (Issue One). That number covers lobbying at all levels of government, not just Congress, but it shows the scale of OpenAI's effort to influence policy. The company is at once the main seller of AI tools to Congress and one of the biggest spenders on lobbying to shape the rules around those same tools.

The broader context here matters. Concentrating purchases on a single vendor creates a practical dependency: if OpenAI raises its prices, changes its terms, or alters how its AI behaves, House offices have few alternatives ready to go. Claude's appearance in the records shows some offices are looking at other options, but with less than one-seventh the number of purchases, that effort is still in its early stages.

The partisan spending gap raises questions of its own. Whether the difference reflects different attitudes toward AI, different office budgets, or simply Democratic staff moving faster on procurement cannot be determined from spending records alone. The data shows where the money went, not why.

What the figures do show clearly is that AI tools have gone from novelty to routine budget item in congressional offices within a very short time. Staff are spending real, if modest, money on them across hundreds of purchases, building them into daily work that covers everything from answering citizens to analyzing legislation.

The total amount, $113,740 across the entire House over a full year, is small by business software standards. A single mid-size company would likely spend more on AI tools. But Congress is not a company. It is the body that will write the laws governing how AI is used, and its staff are now everyday users of the products they may need to regulate, with the leading vendor also among the biggest spenders on influence in Washington.

In this author's view, that is not inherently a problem. But it is a fact worth keeping in mind as the debate over AI rules heats up in the years ahead.