Anthropic’s IPO Valuation Raises A Question: Is AI Worth Its Cost?

AI can help a company produce more. Whether people need, want, or can afford what it produces is another question.

Investor caution surrounding Anthropic’s planned IPO in recent days brings that question into focus. Earlier expectations of a roughly $2 trillion valuation gave way to discussions around $1.5 trillion. This is a reassessment of an anticipated offering, rather than a completed fall in market value, but it reflects growing scrutiny of the economics behind AI’s expansion.

Anthropic’s November 2 filing disclosed approximately $895.4 billion in purchase commitments and leases, mostly for computing and infrastructure, extending into the following two decades. These are not all outstanding debt, and some are conditional. Nevertheless, they represent enormous claims on future resources. Third-quarter revenue reached approximately $17 billion, yet compute and infrastructure expenses rose faster than revenue from the preceding quarter, and the adjusted operating margin narrowed. More importantly, whether the products and services AI enables can create broad social value remains a question.

In the 19th century, industrialization expanded the supply of goods and changed the conditions of everyday life. Railways and steamships moved people and products across distances previously prohibitive in time or cost. Mechanized production made many goods more widely available. These changes brought exploitation and disruption as well as prosperity, but their contribution to society’s productive capacity was tangible.

Computers and the internet subsequently transformed access and coordination. People could communicate across continents instantly, transfer money without visiting a bank, and reach customers beyond their local markets. They not only produced computer hardware and internet infrastructure, but also made activities possible or accessible on a different scale.