deniz.in

Markets

Weather

Loading weather

· via The Verge

Anthropic's record $81.3B IPO cashes in suppliers and tests its safety-first governance

Anthropic raised $81.3 billion in the largest IPO on record, handing Amazon and Google stakes worth a combined $307.9 billion while analysts debate whether AI's circular financing loop can survive contact with public markets.

Anthropic's record $81.3B IPO cashes in suppliers and tests its safety-first governance

The largest IPO on record

Anthropic has completed the biggest initial public offering in history, raising $81.3 billion and overtaking SpaceX as the most valuable debut ever, according to The Verge. TechCrunch reports the shares priced at $83, and that once the deal settles — including a $26 billion private placement from a group of chip and cloud companies — Anthropic is expected to hold roughly $220 billion in cash, cash equivalents and short-term investments.

Gil Luria, head of technology research at D.A. Davidson, told TechCrunch that Anthropic raised about $141 billion in capital overall and will keep around $91.7 billion in net proceeds, giving it more cash on hand than any company has ever held. The two outlets frame the haul differently — The Verge cites the $81.3 billion IPO figure, while TechCrunch's totals fold in the placement — but the scale is not in dispute. TechCrunch also notes the company still runs a quarterly net loss in the tens of billions, even as Luria argues it is no longer burning cash.

Suppliers become winners

The IPO's biggest beneficiaries include Anthropic's own infrastructure suppliers. Per TechCrunch, Amazon's 14.7% stake — built on an $18 billion investment — is now worth upward of $159 billion, and Google's 13.7% stake is worth nearly $149 billion, a combined $307.9 billion at the IPO price. Google has invested $12.3 billion since 2023, with another $30 billion promised if Anthropic hits performance milestones. Nvidia, AMD, Microsoft, Broadcom, Micron and TSMC joined Google in the $26 billion private placement, and all of them except TSMC also hold multi-billion-dollar partnership deals under which Anthropic buys their products.

That arrangement is what critics call circular financing: money invested by suppliers flows straight back to them when Anthropic trains models or serves Claude. The listing makes those positions liquid and, in TechCrunch's reading, shifts the funding burden from conflicted supplier-investors to public shareholders and corporate bonds.

Analysts are split on how fragile that system is. Joey Brookhart, who leads the Tokenomics team at SemiAnalysis, estimates Anthropic needs about $280 billion in cumulative revenue over four years to cover its compute bill — a target he calls achievable given the company's $17 billion in third-quarter revenue, roughly a $68 billion annualized pace. Anthropic has already committed $900 billion, mostly to compute and infrastructure. Luria contrasts Anthropic with CoreWeave, which he says exists only because Nvidia props it up, while Brookhart argues Anthropic's financing has been additive to the ecosystem rather than purely circular.

The governance fine print

The Verge's reporting focuses on a second tension: whether a company founded on safety priorities can keep them as a public entity. Anthropic is registered as a Public Benefit Corporation, and its Long-Term Benefit Trust (LTBT), created in 2023, controls a special Class T share class that can select and remove part of the board. But the details are narrower than the mission statement suggests. The trust picks four of seven directors, one of whom effectively requires founder approval, and removals require 30 days' notice to a governance committee chaired by co-founder Daniela Amodei alongside Reed Hastings and investor Yasmin Razavi. The founders also hold controlling voting rights, meaning shareholders cannot vote them out.

The four listed trustees — former Fed Chair Ben Bernanke, foreign policy analyst Richard Fontaine, physicist Christopher Stubbs and chair Neil Buddy Shah — are not known as AI experts. Ann Lipton, a corporate governance professor at the University of Colorado Law School, told The Verge she doubts the trustees have the technical insight to know when to intervene. Tyler Johnston of The Midas Project noted that Anthropic could blunt a troublesome trustee by simply hiring them, pointing to former trustee Mariano-Florentenio [sic] Cuéllar joining as Chief Global Affairs Officer in August. Unlike OpenAI's old nonprofit board, the LTBT has no authority over day-to-day decisions, according to The Verge's reading of the S-1, leaving a structure closer to Meta or SpaceX than to an independent check.

Why it matters

Anthropic's debut sets the template for every AI lab that follows it into public markets. It is the first large-scale test of whether AI's supplier-funded financing loop can be replaced by ordinary market capital, and whether a safety charter can survive quarterly pressure when founders, not shareholders or trustees, hold nearly all the real power. Both analysts told TechCrunch the biggest risk is competitive, not financial: if OpenAI, Google DeepMind or open-weight models take the lead, Anthropic's valuation collapses while the compute sellers still get paid. As The Verge notes, the constraint on Anthropic's mission has always been that it needs enormous amounts of money — and now the market, not the trust, will be watching.

  • #anthropic
  • #ipo
  • #ai-safety
  • #venture-capital
  • #cloud-computing

Related posts