ChatGPT dominates 90% of congressional AI spending.

United States congressional offices are overwhelmingly choosing OpenAI's ChatGPT as their primary generative AI assistant, according to House disbursement records for the year ending March 31. During this period, House offices, committees, and institutional accounts spent at least $113,740 on dedicated AI tools, with OpenAI securing approximately 90% of that total. Specifically, Congress directed $100,580 toward ChatGPT across 798 separate transactions. Anthropic's Claude occupied a distant second place, capturing $13,160 across 37 transactions.

The spending data also highlights a significant partisan divide in AI adoption. Democratic offices accounted for $54,165 of the total AI purchases, which is more than three times the $15,782 spent by their Republican counterparts. These figures only represent direct, paid transactions for standalone AI tools, excluding free accounts or AI capabilities bundled into broader enterprise software contracts. Staffers are deploying these tools for a variety of daily tasks, including drafting policy memos, summarizing complex legislation, responding to constituent inquiries, preparing hearing materials, analyzing policy research, and writing social media updates.

For AI practitioners and enterprise developers, this lopsided adoption demonstrates that first-mover advantage and brand recognition remain incredibly powerful in the public sector. Even as competitors like Anthropic pitch superior reasoning or larger context windows, ChatGPT's ubiquity makes it the default choice for government workflows. Developers looking to sell to federal clients must recognize that ease of procurement and established trust currently outweigh minor benchmark differences. Furthermore, the partisan gap in spending suggests that go-to-market strategies for government-focused AI startups may need to be tailored differently depending on the political affiliation of the target office.