Protesters hold “Freeze AI Now” signs in front of OpenAI headquarters in San Francisco on October 20, 2026.
Heather Deal | Getty Images
Silicon Valley’s attempts to control artificial intelligence politics have turned into a war with itself.
With just one week to go until the midterm elections, political groups supporting faster AI development and lighter, more uniform rules have spent about $8 million in general election congressional races, compared to about $6.9 million from groups pushing for stronger safeguards and more regulatory leeway for states, according to a CNBC analysis of Federal Election Commission filings through Oct. 20.
In Florida’s 25th Congressional District, this split has led to a direct showdown. Think Big, an affiliate of Leading the Future, a major super PAC network pushing for faster AI development and less regulation, spent $310,000 during the primary to support Democrat Jared Moskowitz, whose district was redistricted after Florida redrawn its congressional maps for the 2026 election. Meanwhile, the super PAC Defending Our Values, which supports stricter security fences and state-level powers, dropped about $857,000 to boost Republican Scott Singer and another $740,000 to keep one of his primary opponents from winning.
Now, Moskowitz and Singer will face off in the general election.
This epitomizes the failure of Silicon Valley’s attempts to build a unified AI political machine. While both sides support AI innovation and agree on the need for guardrails, they remain locked in a costly tug-of-war over which candidate to support and which candidate best fits their desired speed of deployment, regulatory scope, and balance of state power.
Electing more allies could give either side an advantage in shaping AI policy in the next Congress, even if the feud risks weakening Silicon Valley’s influence.
“Everyone supports some form of regulation. It’s just a debate about how much regulation that is,” said Bruce Freed, president and co-founder of the bipartisan Political Accountability Center and former chief investigator on the Senate Banking Committee. “Money is flowing in all directions to control what happens.”
CNBC identified the federal committees that work with each campaign and totaled their own spending supporting or opposing Congressional candidates as reported by the FEC. This analysis excludes spending on state elections and nonprofit advocacy.
Cryptocurrency playbook
This wasn’t the plan. The model was cryptocurrency.
Support for Fairshake and its related crypto-promoting super PACs coinbaseRipple and Andreessen Horowitz spent more than $130 million during the 2024 election cycle, according to FEC filings. According to cryptocurrency publication CoinDesk, candidates backed by the company won 53 out of 58 races, giving them a 91% chance of winning.
AI investors took notice. In August 2025, Leading the Future announced that Andreessen Horowitz, OpenAI President Greg Brockman and his wife Anna, Perplexity & Palantir Co-founder Joe Lonsdale.
The network, co-led by Josh Brust, who worked at FairShake last election cycle, developed the Democratic PAC Think Big and the Republican PAC American Mission to support candidates aligned with opposition to accelerating AI development and a patchwork of state regulations.
The message was clear. If you don’t back out of your AI plans, you risk millions of dollars being spent against you.
“The AI industry announced last year that it was going to come in with a lot of money and repeat the crypto industry playbook to bully lawmakers into submission,” said Sasha Howarth, founder and executive director of the Tech Oversight Project, a nonprofit watchdog group that advocates for stronger regulation of Big Tech and AI.
It worked fine at first. In the first major primary of the cycle in March, three candidates heavily supported by American Mission — Jessica Steinman of Texas, Chris Gover of Texas and Laurie Buckhout of North Carolina — won Republican races after the group spent more than $500,000 in support each, according to FEC filings.
But rival infrastructure was already taking shape. Anthropic committed $20 million to Public First Action in February and another $20 million in July, giving the Safeguard camp a major financial boost. Anthropic says its funds are limited to policy and public education and cannot fund the election of candidates. Still, another funded super PAC associated with Public First Action has spent millions of dollars promoting similar priorities.
The conflict came to light during the Democratic primary for New York’s 12th District. Think Big spent $8.3 million to defeat state Rep. Alex Boaz, who authored the state’s major AI safety law, according to FEC filings. Humanity Backed Jobs and Democracy PAC responded by plowing about $15 million into his support, while allied safety groups spent millions more, according to a CNBC analysis.
The analysis shows that Think Big and Jobs and Democracy have spent a combined $23.3 million, about 40% of each campaign’s major spending nationwide.
In the end, Boaz lost 39% to 35% to Mika Lasher, giving Think Big a tactical victory.
“They helped ignite the fire in the backlash that is currently fueling much of the midterm elections,” Howarth said. “The campaign showed that spending on safety-first candidates can provoke a backlash from the industry itself.”
New York State Congressman Alex Boaz participates in an immigration protest on September 18, 2025.
Eric McGregor | Light Rocket | Getty Images
money flow slows down
The battle dragged on until the general election, but the pace of spending slowed sharply.
Safeguards-focused groups spent $825,000 this fall against two Republican senators who benefited from Future Usher spending during the primaries, according to a CNBC analysis of FEC filings. These two races alone account for more than one-sixth of the Safeguard camp’s general election opposition spending.
According to FEC filings, Guardrail Alliance reported spending $412,500 each on Republican House candidates Jay Feely of Arizona and Aaron Flynt of Montana, both of whom benefited from spending aligned with Leading the Future in the primaries. A CNBC analysis also reported about $412,000 spent on Democratic Rep. Jimmy Gomez of California, whom Leading the Future endorsed, but no aid spending was reported.
“Big AI groups still believe their money can help influence elections, even if it’s just for a couple in November,” Howarth said.
But the overall flow of AI campaign finance has slowed since the primaries.
The Leading the Future network reported major independent spending of about $26 million, compared to about $8 million in the general election. Groups focused on safeguards reported about $32.1 million in primary spending and about $6.9 million in general election spending. Those groups reported about $5.5 million more in runoff and special election spending.
About 74% of the roughly $78.5 million spent on federal candidates this cycle went to primaries, according to a CNBC analysis.
Part of this decline reflects the smaller size of the November race. But this slowdown has come at the same time as AI has become more difficult to sell politically.
According to a new CNBC National Economic Survey, 52% of registered voters say companies should pause development of advanced AI until Congress passes regulations, compared with 32% who support moving forward. 63% said they would oppose having an AI data center within 10 miles of their home.
“If you weigh the risks and rewards at this point, the backlash is too great,” said Fried, of the Center for Political Accountability.
“AI executives and groups spent a lot of money on primaries, so they likely achieved their goals in most campaigns. There’s no big downside to scaling back at this point,” Fried said.
Some major donors have gone further.
The Brockmans have donated $25 million to Leading the Future. But as OpenAI employees spoke out against the group’s tactics, OpenAI’s chief strategy officer acknowledged internally that the company was suffering a “reputational hit,” Semaphore reported.
The Brockmans have since abandoned plans to donate an additional $25 million, according to The New York Times.
Meta also withdrew. Political organizations it funded returned $25 million this summer, according to IRS and state filings. The funding included $22 million from the American Technology Excellence Project, which promotes AI, and $3 million from operations in California. The refund reduced a planned $65 million state-level political push by nearly 40% to about $40 million. The Times first reported the refund.
“Obviously something isn’t working,” OpenAI CEO Sam Altman told Politico earlier this month. “I don’t think spending money on politics makes people feel better.”
