The Anthropic Investor Calling His Own Portfolio's Lobbying an Oligopoly Play — Right Before a $2T IPO

On Friday September 25 at Reuters Momentum AI Austin, Palantir co-founder and 8VC founder Joe Lonsdale — who counts himself as an Anthropic investor — accused OpenAI and Anthropic of manufacturing fear about existential AI risk to influence Washington. 'I think that what OpenAI and Anthropic are pushing right now is very dangerous,' Lonsdale said, 'and we don't want an oligopoly that controls all of our policy here with the government.' He added: 'Safety is a real thing, but fear and regulation — we've got to be really careful how we respond to those things.' He stopped short of disowning the company — he praised Anthropic's talent and called it a leader in the AI race — but pivots from ally to accuser on Anthropic's own lobbying play, a posture that no other major Anthropic backer has taken publicly.

Lonsdale's 'oligopoly' line is not abstract. On September 15 OpenAI confirmed weeks of closed-door talks with Anthropic and Google DeepMind, and The Information reports the three are finalising a voluntary body — tentatively named the Standards Authority for Frontier AI (SAFA) — to launch by late 2026 or early 2027, with Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman publicly backing a slowdown of frontier development. Two days earlier, on September 13, former Trump AI czar David Sacks wrote on X: 'Stop pretending antitrust law has to be suspended so you can form a cartel. Stop pretending you need a regulatory approval process.' Lonsdale is now echoing that critique from inside the Anthropic cap table, which converts a right-wing policy attack into an investor-level dispute over Anthropic's IPO narrative.

Whatever the merits of Lonsdale's argument, the antitrust frame collapses Microsoft's two biggest AI bets into one investigation. After the October 28, 2025 restructure Microsoft holds a 27% stake in OpenAI Group PBC (valued at ~$135B at the time), gains access to OpenAI models through 2032, and capped its OpenAI revenue share at $38B through 2030. The same company also committed $5B of equity to Anthropic in November 2025 alongside $10B from NVIDIA, against Anthropic's $30B Azure compute commitment. The duopoly frame therefore puts Microsoft at the center of any 'safety cartel' story, even though the antitrust target would more naturally be the three frontier labs themselves.

Anthropic is now pacing toward a mid-October 2026 IPO at a ~$2T valuation, with a raise as large as $100B. The math only works on a benign regulatory landing: a $2T market cap on ~$65B of run-rate revenue implies a ~30x sales multiple — already a stretch against any public software peer, and roughly twice Microsoft's trailing EV/Sales of ~11.9x. If antitrust scrutiny of a safety cartel reopens the legislative window — Senator Van Hollen's September 10 letter to Sam Altman is the first concrete data point — that multiple compresses quickly. Lonsdale's intervention converts a story about safety into a story about regulatory capture, exactly the framing IPO investors hate most.

Lonsdale's intervention doesn't stand alone. The White House has been moving against safety-led regulation for almost a year: a December 11, 2025 executive order ('Ensuring a National Policy Framework for Artificial Intelligence') establishes an AI Litigation Task Force under the Attorney General and pre-empts state laws; Trump publicly referred to those calling for AI regulation as 'negative forces' in mid-September; and David Sacks — who ran Trump's AI and crypto portfolio until March 2026 — publicly rejected the antitrust-waiver route on September 13. The administration's preferred alternative is product liability enforced through the courts, not pre-approval. The practical implication for Anthropic is that the regulatory tail risk in its $2T IPO is no longer purely theoretical — the White House, the Trump-aligned investor class, and now a major Anthropic backer all want the safety coalition to lose.