OpenAI builds lobbying muscle

OpenAI is no longer just sending engineers and lawyers into government meetings. It has been building a full Washington operation, with federal lobbying filings, Washington job postings, and a state-policy team aimed at the rules that will decide how artificial intelligence products are sold, audited, and used by government buyers. (opensecrets.org, openai.com, jobs.ashbyhq.com) The money moved fast. OpenAI spent $260,000 on Capitol Hill in 2023, then $1.76 million in 2024, and OpenSecrets says it had already spent $2.1 million in 2025 at the time of its latest tally. (technologyreview.com, opensecrets.org) Its federal disclosure forms show this is not a trade-group side project. OpenAI OpCo filed under its own name for 2025 reporting, which is what companies do when they want direct control over the message instead of borrowing someone else’s megaphone. (lda.senate.gov) The hiring tells the same story. OpenAI advertised a United States Executive Branch lead in Washington with pay listed at $171,000 to $280,000, and it separately posted for a State Government Affairs lead to manage relationships with elected officials, agencies, and local governments across multiple states. (jobs.ashbyhq.com, jobs.ashbyhq.com) This is happening because artificial intelligence policy is no longer one fight in Congress. It is dozens of fights at once over watermarking, child safety, workplace rules, government procurement, model testing, and whether states can write their own laws at all. (politico.com, watech.wa.gov) Washington state shows why companies are staffing up locally. On March 25, 2026, Governor Bob Ferguson signed bills requiring traceable labels for some artificial-intelligence-generated content and repeated disclosures from chatbot services that users are talking to software, not a person. (opb.org) Washington is also building a bigger state framework behind those laws. Its Artificial Intelligence Task Force said in a December 2025 interim report that the federal government’s “hands-off approach” had left a regulatory gap, and it recommended transparency, workplace rules, health-care accountability, and law-enforcement disclosure standards. (watech.wa.gov) At the same time, OpenAI is trying to become a government supplier, not just a company regulated by government. It launched OpenAI for Government on June 16, 2025, and said the program would bring its tools to public servants across the United States. (openai.com) That sales push got real money behind it. In August 2025, the General Services Administration said it had struck a OneGov deal to make ChatGPT available government-wide through the Multiple Award Schedule, with pricing described as $1 per agency. (gsa.gov) So OpenAI is now playing both sides of the same table. It is lobbying the people writing the rules while also courting the agencies buying the tools, which means every fight over disclosure, safety testing, procurement language, and state enforcement now hits its business twice. (openai.com, gsa.gov, politico.com) And OpenAI is not doing this alone. Bloomberg Law reported that artificial-intelligence lobbying spending in the last quarter of 2025 was up 370% from the quarter when ChatGPT launched in late 2022, while Politico reported that OpenAI, Anthropic, Nvidia, and Scale AI were all expanding their Washington presence. (news.bloomberglaw.com, politico.com) The old image of artificial intelligence companies was a handful of labs asking to be regulated carefully. The new image is a full campaign map: Capitol Hill, statehouses, procurement offices, and agency budgets, all at once. (technologyreview.com, politico.com, opb.org)