Anthropic IPO: 5 points to know behind its $2 trillion bet, $42 billion loss and warning of existential AI risks to humanity

Anthropic IPO: Anthropic’s potential IPO is starting to look very different from a normal technology listing.

The AI company is growing at extraordinary speed. Its revenue has surged into the billions, while its spending has reached another scale altogether. A leaked draft of its IPO prospectus, reported by Reuters, gives investors an early look at the numbers behind that expansion.

The picture is complicated. Anthropic is showing explosive sales growth, but it is also preparing for enormous future infrastructure costs.


Here are five points to know about the Anthropic IPO.

1. Anthropic’s revenue jumped more than 1,000% in one year

The easiest number to notice is also one of the most striking.

Anthropic generated about $4.59 billion in revenue in 2025, according to the leaked prospectus reported by Reuters. That was up from roughly $386 million in 2024, an increase of about 1,088%.

For a company operating in one of the most expensive areas of technology, that kind of growth helps explain the excitement surrounding its potential public offering.

The momentum appears to have continued into 2026. Anthropic reported $4.73 billion in revenue in the first quarter, according to CNBC, before revenue climbed to about $11.5 billion in the second quarter.

2. The $42 billion loss changes the picture

The company reported a net loss of about $42 billion in 2025, compared with a loss of roughly $8.31 billion in 2024. Its operating expenses reached around $13 billion, while its operating loss exceeded $8 billion.

That is an unusual combination: revenue rising rapidly while losses remain enormous.

AI companies face costs that many conventional software businesses do not. Training and operating advanced models requires large amounts of computing power. Anthropic also needs engineers, researchers, data-center capacity and cloud infrastructure.

3. The $518 billion infrastructure commitment may be the biggest number to watch

There is a figure in the leaked prospectus that puts Anthropic’s ambitions into perspective: about $518 billion in commitments for cloud services, computing and infrastructure.

That is vastly larger than the company's current annual revenue.

The number matters because Anthropic cannot scale its models without the physical and computing infrastructure behind them. Every new wave of AI demand requires more processing capacity, storage and data-center resources.

It also connects Anthropic to a much wider group of companies.

Cloud providers, chipmakers, data-center operators and other infrastructure businesses all benefit when AI companies increase spending. Anthropic’s expansion is therefore not happening in isolation.

The reverse is also true. If demand for its models fails to match its infrastructure plans, the economics of that spending become much harder to justify.

4. A $2 trillion valuation would require investors to believe the future is much bigger

Anthropic was valued at about $965 billion in its May funding round.

The reported target of roughly $2 trillion for an IPO would therefore represent a dramatic jump from its latest private valuation.

That valuation cannot be explained by today's profits. Anthropic is still reporting enormous losses.

Instead, investors would have to place substantial value on what the company could generate in the years ahead.

That is common in fast-growing technology businesses, but Anthropic's numbers make the gap especially visible. The company has rapidly expanding revenue, yet its infrastructure requirements are expanding alongside it.

5. Anthropic has two very different risks hiding inside its IPO story

The prospectus reportedly devotes extensive space to risks associated with Anthropic's technology.

Among them are warnings about increasingly capable AI systems and the possibility of unpredictable behavior, including scenarios the company describes as posing existential risks to humanity.

There is also a much more conventional business risk.

About a quarter of Anthropic's revenue reportedly comes from just two customers. That means a significant share of the company's sales is concentrated among a very small number of buyers.

The two risks are very different, but they point to the same problem facing Anthropic as a public company.

Its future depends on maintaining rapid growth while managing a technology that is expensive to develop and operate. It also needs customers to keep increasing their use of its models.

That is what makes the potential IPO unusual.

Anthropic is not coming to market as a mature software company with predictable earnings. It is presenting investors with a much larger bet: that demand for advanced AI will continue growing fast enough to justify extraordinary infrastructure spending and eventually turn that spending into sustainable profits.