Anthropic's IPO prospectus reveals massive costs and limited investor control

The scale of the AI ambition

Anthropic is preparing for an initial public offering that could redefine the scale of the tech industry. The company's preliminary prospectus suggests it may raise about $100 billion, which would put its estimated valuation as high as $2 trillion. This would make it one of the largest stock issues in recent history, surpassing the SpaceX IPO in size.

To achieve its goals, Anthropic has entered into several non-cancellable agreements for computing resources and chips. These commitments total roughly $500 billion, including significant payments to Amazon, Google, Microsoft, and SpaceX.AI. Because these deals are non-cancellable, the company remains on the hook for these costs regardless of whether it actually uses the hardware.

The financial reality of this growth is visible in the company's recent performance. While revenue jumped to $16.3 billion in the first six months of this year, operating expenses also surged to nearly $9 billion. This reflects the intense competition for the infrastructure needed to train and deploy advanced artificial intelligence models.

Risk and the 'Company Purpose'

The prospectus places a heavy emphasis on what Anthropic calls its "Company Purpose." This principle, which appears 42 times in the document, states that the company intends to develop AI for the benefit of humanity. However, the filing also includes a sobering disclaimer for potential investors regarding the technology's unpredictable nature.

Anthropic explicitly acknowledges that it cannot guarantee that its work will result in a positive impact. The document warns that highly sophisticated AI models might create devastating or terminal threats to the world if they are not handled with care. This creates a paradox where the company's stated mission is paired with a formal admission that the technology might actually be dangerous.

Critics point out that while these goals are noble, they are not legally binding guarantees. The prospectus serves as a warning that the company is essentially playing with fire to build its multi-billion dollar business. It highlights the gap between the idealistic goals of the founders and the high-risk reality of the current AI development cycle.

A fortress of founder control

One of the most striking features of the Anthropic filing is its unique five-class share structure. While the Class A common shares will be traded on the Nasdaq, they will never have more than 49.9% of the voting power. This means that regular investors will not have any say in how the company is managed or how its policies are decided.

The majority of the voting power, specifically 50.1%, is held by a single Class F share. This share is controlled jointly by the company's seven co-founders, including CEO Dario Amodei and his sister Daniela. This arrangement is designed to remain in place until the number of remaining co-founders drops below three, ensuring the founders maintain ultimate authority.

There is also a Class T share held by an independent Long-Term Benefit Trust. This trust is managed by specialists in public policy and AI safety and has the power to elect a majority of the board. While this is intended to protect the company's mission, the primary control remains firmly in the hands of the original founders, making their leadership effectively impregnable to outside pressure.

The path to profitability

For Anthropic to succeed as a public company, its revenues must significantly outpace its massive ongoing expenses. Currently, a substantial portion of its revenue comes from a limited number of customers, with its five largest clients providing about 25% of its income. Many of these customers are not bound by long-term contracts and can stop spending at any time.

This creates a precarious situation if corporate demand for AI begins to cool or if competition intensifies. Some economists have already noted that, outside of the tech sector, AI has yet to deliver a clear financial payoff for many corporations. If non-tech companies pull back on their spending as fees rise, Anthropic could face significant financial pressure.

The company's future depends on its ability to maintain its current pace of growth while managing a mountain of debt and infrastructure costs. Investors are essentially betting that the demand for advanced AI will remain high enough to cover the $500 billion in commitments. If that demand wanes, the company's high valuation and massive overhead could become a significant liability.