Anthropic Has Questions To Answer Before IPO, Says Economic Liberties

November 10, 2026 Press Release

Washington, D.C. —Today, the American Economic Liberties Project released a new memo, “What the SEC and Investors Should Ask as Frontier AI Companies Go Public,” laying out questions regulators, investors, and pension fund fiduciaries should be asking ahead of Anthropic’s planned IPO, given the risk it poses not just to Wall Street and Silicon Valley, but more concerningly the entire U.S. economy, including the retirement savings of workers across the nation.  

The memo works as an accountability blueprint not only for Athropic’s IPO — which is expected to list as soon as next week — but also for similar offerings that will likely be made by frontier AI firms after. IPOs of this size are typically pulled into the major indexes within days of listings, and the risks will land in the 401(k)s and pensions of millions of working Americans who never chose to invest in these companies. The memo closes with a template of specific questions that fiduciaries, SEC staff, members of Congress, state attorneys general and reporters can put to these companies directly.  

“Frontier AI IPOs will expose countless working people to enormous financial risks through their retirement savings, regardless of their decision to invest in these companies or not,” said Colin Dunkley, Legal Analyst at the American Economic Liberties Project. “The SEC has the authority to demand full disclosure of hidden liabilities, conflicted financial relationships, and accelerating litigation risks before the company goes public. It should use that authority now rather than wait for financial disaster.”

“The mountain of unanswered legal and financial questions casts serious doubt on how Anthropic’s IPO is being priced,” said Laurel Kilgour, Research Manager at the American Economic Liberties Project. “Wall Street is already, wisely, revising this offering’s valuation downward, but a lower number isn’t the same as clear answers. Pension fund fiduciaries and regulators have an obligation to get real answers rather than fall for the hype of these circularly financed frontier labs.” 

The memo identified key risks:

  • Circular financing. Google, Amazon, Microsoft, NVIDIA, AMD, and Broadcom simultaneously serve as Anthropic’s biggest investors, lenders, rivals, customers, and suppliers, making it difficult to discern how much genuine, independent demand exists for Anthropic’s products. 
  • Massive obligations. The S-1’s disclosed commitments add up to roughly $855 billion, mostly noncancelable, against $10.8 billion in contractually obligated customer revenue; a roughly 79-to-1 mismatch. The ratio that brought down WeWork was 11.8 to 1.
  • Hidden debt. Third-party leasing structures keeps much of Anthropic’s leverage off its balance sheets, including roughly $71 billion in off-balance-sheet chip-leasing debt taken on in a single 60-day stretch. A major index fund has reportedly told investors Anthropic is likely to receive a junk credit rating.
  • Mounting legal exposure. Anthropic faces litigation from multiple angles at once — wrongful death, copyright, defamation, and personal injury — risking billions in damages. 
  • Insular corporate governance. Founders hold voting control, a Long-Term Benefit Trust elects a majority of directors, and reported lock-up terms would let insiders begin selling on day one rather than observing customary 180-day lock up. 
  • Catastrophic risk, disclosed in their own words. Anthropic’s own S-1 warns that its systems could “pose catastrophic or existential risks to humanity.” These risks run through decisions made by humans, and investors are entitled to know how those decisions are made.

Some of the questions the memo says any frontier AI company should answer before going public

  • If insiders can sell shares before customary 180-day lock-up expires, how many shares and on what schedule?
  • What are the company’s total non-cancelable commitments, broken out by counterparty and by year, and what happens to its liquidity if demand falls? 
  • What share of the company’s capital comes from firms that are also its suppliers, customers, or competitors? 
  • For Anthropic specifically: What internal estimates exist of the probability and magnitude of the catastrophic risks the company itself describes, and has the testing behind its safety claims been independently verified?
  • For Anthropic specifically: What portion of its reported $600B in compute commitments is non-cancelable, and on what schedule? Which entities actually hold its off-balance sheet leased chips, and what make-whole obligations does Anthropic owe them? 

Already, leaders, such as New York City ‘s Comptroller, are asking similar questions, urging Anthropic to pause its IPO until it addresses the risks it has itself identified. 

As the memo notes: Fiduciaries and investors must press for fuller disclosure during the comment window, the SEC must use its review powers to ensure the S-1 is complete and accurate. In addition, state attorneys general retain independent antifraud authority that issuer arbitration clauses and class-action waivers cannot block. 

Read the full memo here.

Learn more about Economic Liberties here.