OpenAI plans to host the grand opening of its first lobbying office in Washington, called the Workshop, on Wednesday. The artificial intelligence start-up has said it created the space — part lab, part showroom — just blocks from the White House to better work alongside lawmakers.
The office is part of OpenAI’s increasingly aggressive push to sway A.I. policy. The company has lobbied for the proliferation of data centers, which are needed to power the technology, and pressed to freely use copyrighted material. It spent $1 million on federal lobbying in the first quarter, double the amount a year earlier, according to congressional disclosures.
Just blocks away, an A.I. rival, Anthropic, opened its first office in Washington in April, as it battled with the Pentagon over the use of its technology. It has hired six lobbying firms in recent months and increased its spending on Washington lobbying tenfold to $3 million last year, according to disclosures.
The activity by A.I. companies in the nation’s capital has reached a fever pitch, as they open offices, hire lobbyists and host large conferences to pitch policy ideas and promote their technologies.
A quarter of the 13,000 federal lobbyists in Washington are involved in A.I. issues, up from 11 percent in 2023, according to an analysis of congressional disclosures by Public Citizen, a nonprofit watchdog group. Meta, Nvidia and Alphabet, Google’s parent company, spent a combined $47.8 million on federal lobbying last year, up 22 percent from 2024, according to Senate disclosures. Meta and Alphabet were top corporate spenders.
“We’re seeing an unprecedented deluge of money being poured by A.I. companies into lobbying in order to protect their bottom lines and their images at a time when Americans are very anxious about the technology,” said Isabel Sunderland, the policy lead for tech at Issue One, a nonprofit government accountability group.
Winning over federal lawmakers has taken on new urgency as states have introduced dozens of bills this year to put guardrails around A.I. The Trump administration — which once said American companies should have mostly free rein to develop the technology — is also considering the introduction of government oversight over new A.I. models.
At stake is the future of A.I. development. OpenAI, Meta and Google have pushed for little to no regulation, saying restrictions would harm their chances in an A.I. race with China. Anthropic and others have supported new laws, pointing to the technology’s potential dangers.
A.I. faces public skepticism in the lead-up to November’s midterm elections. Voters have expressed concerned about power-guzzling data centers that power A.I. and rising electricity costs, alongside fears the technology could disrupt the economy.
Parents’ groups have also sounded the alarm about children’s interactions with chatbots, after which some teenagers took their own lives. In a recent NBC News poll, 57 percent of registered voters said A.I.’s risks outweighed its benefits, compared with 34 percent who said the opposite.
Most A.I. companies said they were open to legislation that promoted innovation and the technology’s development.
“This is a general-purpose technology on the scale of the wheel, printing press, combustion engine, electricity,” said Chris Lehane, chief global affairs officer for OpenAI. “We at OpenAI have felt for some time now that the conversation on policy solutions and policy needs need to be as transformative and big as the underlying tech itself.”
(The New York Times has sued OpenAI and Microsoft, claiming copyright infringement of news content related to A.I. systems. The two companies have denied the suit’s claims.)
“We’re advocating for policymakers to come together on federal legislation that promotes American leadership in A.I.,” Julie McAlister, a Google spokeswoman, said.
Alongside A.I. companies, communications firms, trade groups and think tanks have bulked up their presence in Washington on both sides of the A.I. issue.
Last year, the Facebook co-founder Dustin Moskovitz’s philanthropy, Coefficient Giving, funded a new communications and lobbying group pushing for A.I. regulation. The group, the Alliance for Secure A.I., wants strong chatbot regulations to protect youths. It is also aiming for more safety oversight of A.I. models and has opposed efforts by President Trump and some federal lawmakers to pre-empt states from creating A.I. laws.
The group’s chief executive, Brad Steinhauser, a former Tea Party leader, has lobbied in Congress and met with Texas legislators, including State Senator Angela Paxton, to push for child safety and other measures. He has also appeared on podcasts and other media.
“I’m going on ‘Bannon’s War Room,’ NPR or The New York Times — anywhere and everywhere to get the message out that policymakers need to move fast to protect citizens,” Mr. Steinhauser said.
OpenAI and Anthropic have been the most active. In September, Anthropic made its official Washington lobbying debut by holding a daylong event at the city’s Union Station. Dario Amodei and Jack Clark, who are among Anthropic’s founders, welcomed hundreds of policymakers and Trump administration officials to showcase the company’s technology.
“We’ve always advocated for basic transparency requirements around models,” Mr. Amodei said in at the event. “Many of the risks we’re most worried about are approaching us. They’re on the horizon.”
Anthropic tripled its policy staff last year and plans to triple that number again this year. It named its first head of lobbying, Anthony Cimino, in January.
In February, the company became embroiled in a legal dispute with the Pentagon over the use of A.I. in warfare and was designated a “supply-chain risk to national security.” It hired a Trump-connected lobbying group, Ballard Partners, in March to support its case with the White House.
Anthropic then opened its office in Washington last month, with large event spaces to show its technology to regulators and discuss the effects of A.I. on national security, the economy and safety.
Last month, Anthropic also released a new A.I. model, Mythos, which it said was so powerful at identifying security vulnerabilities in software that it could lead to a cybersecurity “reckoning.” That helped start discussions within the White House about government oversight of A.I. models.
Anthropic and OpenAI have had regular talks with the White House about a potential executive order on model testing, the companies said.
“We’re focused on developing A.I. safely and ensuring that America leads in A.I., and making that happen requires close partnership between industry and government,” said Sarah Heck, Anthropic’s head of policy.
After OpenAI opens its Washington office — in the Gallup Building, a former Masonic temple — it has planned a series of inaugural events there.
The sessions include training local high schoolers and senior adults on how to use A.I. The company will then begin hosting policy discussions with lawmakers and Trump administration officials at the space.
“What’s that line from ‘Hamilton’?” OpenAI’s Mr. Lehane said. “It’ll be the room where it happens.”
