PBC’s public listing prospectus warned investors of “existential risks to humanity.”Why? Being long or short the artificial intelligence lab’s shares will not matter to investors wiped out by Skynet, the fictional AI of the movies that seems to some a vision of the future.

Prospectuses are written to protect issuers, not to help those purchasing shares in an initial public offering. Do Anthropic’s lawyers anticipate being sued in the afterlife by investors who claim they would not haved to the IPO if they had known AI might take over the world?

No. Anthropic’s lawyers anticipate being sued by shareholders angry that the company gave away profits to prevent existential threats to humanity: slowing down research and model release, lobbying for restrictions on AI use, even closing down the business.

Anthropic is a public-benefit corporation built so that management can sacrifice profit for safety. That’s what investors should think about when reading its prospectus, not whether HAL 9000 will open the pod bay doors for Dave.

The other possible reason for the disclosure is as a stage magician's flash and smoke, so you don't watch his hands. The end-of-the-world disclosure is the flash. The lack of useful information for valuation in the prospectus is the hands.

This is not Anthropic’s fault. The ideal prospectus helps investors judge the probability distribution of net revenue that the business will generate over the foreseeable future and includes the legal and financial information necessary to estimate a fair stock price from that net revenue. The rules were written from the perspective of a 19th century business borrowing money to build a factory with a 40-year expected life, making tangible products to be sold in a stable market.

Let’s start with the balance sheet. The tangible assets Anthropic owns, net of its loans and bonds, are a rounding error in any valuation. But it does reportedly have $518 billion of commitments to buy computing resources, $414 billion of which are noncancelable. That sounds a lot like debt used to buy an asset likely to lose value if Anthropic is not doing well.

It’s like an employee in a one-employer town taking out a mortgage to buy a house that will crash in value the day he’s swept up in a wave of job cuts. Actually, it’s worse, since compute resources are a far more volatile and rapidly depreciating asset than residential real estate.

Investors will want to know the precise terms of the commitments and how they will bite in good and bad states of the world for AI. Some of the suppliers are likely shareholders in Anthropic, or customers or distribution partners. That matters a lot for projections.

It’s also important to distinguish capacity to be used to support ongoing revenue (expenses in economic terms) versus future development (investment). None of that will be found in the financial statements. It’s possible some useful information will be in the notes, but I’d be surprised if there weren’t major gaps, some due to Securities and Exchange Commission rules, others due to Anthropic protecting business information.

Equity is the remaining balance sheet item, and it’s a big deal for shareholders. Anthropic had a $34 billion charge in 2025, against less than $5 billion of revenue, for the increase in value of financing instruments that could eventually turn into shares. And that’s only some of the equity Anthropic has given away with options and convertibles.

The prospectus will list the contractual dilution of shareholder gains, but investors will have to guess about the potentially larger new option grants necessary to attract and retain talent, and to negotiate partnerships, for Anthropic to remain a global AI leader.

On the income statement side, Anthropic is reported to have shown private investors projections supporting $190 billion to $200 billion of 2028 revenue, but projections get no safe harbor, so they will not be in the prospectus.

Even if it could be included, that revenue is not comparable to the 19th century factory that could project modest steady revenue growth over a decade or more. The AI services represented by that hypothetical $200 billion haven’t been invented yet, nor have many of the uses to which they will be put. And whatever is sold and used in 2028 will be only distantly related to what is sold and used in 2030.

What Anthropic will offer to the public is not common stock in the sense accountants can describe. It is not a residual claim on the economic value of an enterprise after liabilities have been satisfied.

The entity allows managers to sacrifice investor wealth for the public good. Public money is earmarked to pay bills already undertaken. Shareholders get a seat at a negotiating table that will oversee operations and divide up future net cash flows. Private venture capital investors are familiar with this kind of deal, but regulators and accountants have not figured out how to describe it fairly to public retail investors.

We’ve seen this movie before in the 1990s internet bubble. Financial statements and regulatory disclosure divorced from business reality. The reforms that followed patched up specific problems, but they did not change the underlying model of a firm with legal debt, tangible assets and a long-established system to distribute residual cash among employees and different classes of investors.

The public will be offered something real and possibly very valuable. It just won't be what the prospectus says it is and that will make valuing the stock at IPO anyone’s guess.