SEC Urged to Block Anthropic IPO Over $900B Debt Risks

Anthropic's IPO Faces Scrutiny Over Massive Hidden Obligations

Anthropic's planned initial public offering (IPO) has ignited a fierce debate over financial transparency and investor protection. The AI company, known for its Claude models, filed a registration statement with the SEC in late October 2026, aiming for a November market debut. But a new analysis from Matt Scherer, a fellow at the Open Markets Institute, argues that the SEC should block the IPO until Anthropic provides audited financials and full disclosure of its spending commitments.

Scherer's report, co-authored with Maya Jenkins, highlights that Anthropic's prospectus reveals nearly $900 billion in non-cancelable spending commitments, most stretching into the 2030s. At least $582 billion is owed to just five companies—Amazon, AMD, Broadcom, Google, and Microsoft—for compute infrastructure. Against this, Anthropic holds $96 billion in cash reserves, a gap that Scherer calls "unprecedented for a company making its stock market debut."

Unaudited Financials and Slowing Growth Raise Red Flags

A key concern is that Anthropic's 2026 financial results have not been formally audited. The company's third-quarter figures suggest revenue growth has "slowed to a crawl," according to industry analysts. This is particularly troubling because the company is relying on extraordinary future growth to meet its obligations. The lack of audited statements means investors cannot verify the accuracy of the revenue claims, which Scherer argues violates the SEC's requirement that registration statements not be "on its face incomplete or inaccurate in any material respect."

Compounding this, Anthropic's most significant future obligations are relegated to footnotes rather than the main disclosure. The company has also continued to announce new multi-billion-dollar contracts—such as a recent data center deal with Nscale—even as the IPO approaches, potentially pushing total commitments past $1 trillion.

Market Forces Squeeze AI Margins

Even if Anthropic's financials were audited, the broader AI market presents daunting challenges. Demand for generative AI is uncertain. Multiple studies, including from Deloitte and Apollo, show that most businesses are not seeing positive returns on AI investments. A survey found that the number of Americans using AI at work fell from 46% in 2025 to 33% in 2026. Anthropic's much-hyped Fable 5 frontier model received a "tepid" commercial reception this summer, according to the Financial Times.

Competition is also intense. Anthropic is part of a crowded field including Google, Meta, Microsoft, OpenAI, and SpaceX, all offering models of similar quality. A price war with OpenAI has been ongoing since summer, and Chinese open-weights models add further downward pressure. As a result, Anthropic cannot easily raise prices without losing customers.

Cost Pressures and External Constraints

On the cost side, the infrastructure needed for AI is staggering. A single 1-gigawatt data center costs an estimated $31 billion, according to Bloomberg. Anthropic has committed to 15 GW of capacity and estimates it will need 20-25 GW by 2028. Training models requires tens of thousands of chips, massive energy, and cooling—each query adds to operational costs.

External factors are also piling on. Community backlash against data centers has scuttled or delayed projects across the U.S., limiting capacity. Supply-chain disruptions due to the Iran War and rising energy costs could drive expenses higher. Anthropic's own S-1 filing acknowledges risks from insufficient electrical power and the possibility that on-site natural gas generation becomes more expensive with demand.

Why the SEC Should Step In

Scherer argues that the SEC has the legal authority to delay or stop the IPO under Section 8 of the Securities Act, which allows the agency to act if a registration statement is incomplete or inaccurate. He points to the unaudited financials, the footnoted obligations, and the ongoing announcements of new commitments as evidence that Anthropic's filing is materially misleading.

The stakes are high. If Anthropic goes public, its shares would likely be included in index funds and retirement accounts, meaning workers and retirees would bear losses if the company falters. Given that the stock market is already historically overvalued and concentrated in a few tech giants, adding a highly leveraged AI company could increase systemic risk. Scherer concludes that the SEC should block the IPO until Anthropic provides two consecutive quarters of audited financials and full disclosure of all spending commitments.

Community and Expert Reactions

The proposal has sparked debate. On Hacker News, commenters are skeptical that the SEC would or should block an IPO on these grounds, arguing that the agency's role is not to judge business viability. However, some worry about the exposure of pension funds and index funds to such a risky offering. The controversy underscores a broader tension between innovation and investor protection in the AI boom.

What's Next

Anthropic has not publicly responded to the criticism. The SEC has not indicated whether it will review the filing more deeply. As the IPO date approaches, the pressure on regulators to act will likely intensify. For now, investors must weigh the allure of AI's potential against the sobering math of Anthropic's balance sheet.