Ex-Andreessen Horowitz Partner Blasts AI Industry's $125M Political Influence Campaign

John O'Farrell, the first outside general partner hired at Andreessen Horowitz — who resigned from the firm in May 2025 — is publicly breaking with his former colleagues over what he describes as a coordinated, nine-figure campaign to neutralize AI regulation through electoral spending. The critique is notable precisely because of its source: an insider who helped build one of Silicon Valley's most powerful venture firms is now warning that the same people who once championed technology as a democratizing force are using their wealth to foreclose democratic debate.

The central target of O'Farrell's criticism is a PAC called Leading the Future, which has raised over $125 million. Its contributors include OpenAI co-founder and president Greg Brockman, Palantir co-founder Joe Lonsdale, the AI company Perplexity, and Andreessen Horowitz itself. O'Farrell argues the PAC isn't making a principled case for any particular AI policy vision — it's sending a message to legislators: engage seriously with AI regulation and face electoral consequences. The PAC's first major target is Alex Bores, a New York State assemblyman running for Congress who co-sponsored state-level AI legislation. Leading the Future has spent $6 million on ads against Bores, most of which don't even mention AI, instead focusing on unrelated personal attacks that Bores has disputed.

O'Farrell draws a direct line to the crypto industry's 2024 playbook, noting that Andreessen Horowitz contributed heavily to Fairshake, a crypto PAC that spent tens of millions helping defeat pro-regulation candidates and elect industry-friendly ones — effectively neutering federal crypto oversight. He sees the AI PAC as the same model scaled up. A counter-PAC, Public First Action — backed by executives from Anthropic — has responded with roughly $3 million to support Bores, part of $6 million across four pro-regulation PACs. But O'Farrell rejects that approach too, arguing that large-scale political spending from any direction corrupts the democratic process.

His proposed alternative is substantive: redirect those hundreds of millions into university boot camps to educate legislators and regulators, fund AI-powered improvements to public services in underserved communities, endow policy institutes tackling job displacement and economic distribution, and launch a genuine moonshot to use AI to cure cancer. He argues the industry's fear of bad regulation is understandable but self-defeating — silencing debate produces worse policy, not better. O'Farrell discloses he may contribute his own money to efforts that expose the AI lobby's political influence campaigns, while acknowledging he still holds significant investment positions in Andreessen Horowitz funds, including AI and crypto.