Anthropic’s business plan: Build God or die trying
Anthropic’s draft S-1 filing suggests that its strong sales are helping fund the enormous computing costs needed to develop new AI models. The company’s revenue growth is impressive, but the article argues that its business depends on continued rapid expansion and increasingly capable models, making its prospects uncertain if that progress slows.
Revenue rose fourteenfold to $16.3 billion in the first half of 2026 compared with the same period in 2025; adjusted income was reported at $1.62 billion after excluding an $80 billion accounting loss and other charges. Around three-quarters of customers spent at least $100,000 a year, and net revenue retention was 900 per cent. The article also points to pressure from rival AI firms and open-source models, plus long-term spending commitments extending to 2045; the supplied text ends before detailing those commitments.
In short
- Anthropic’s revenue grew fourteenfold to $16.3 billion.
- Its AI business is profitable on an adjusted basis.
- Continued growth and heavy computing costs remain key risks.
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Anthropic’s rapid revenue growth, high customer spending and strong retention suggest that businesses see substantial value in its models, giving it a plausible way to finance the computing needed for further development. Continued investment could strengthen its products and position in a competitive market, while long-term commitments may secure the infrastructure required to meet demand.
The case against
The company’s results may depend on a fast pace of model improvement and expansion that cannot be assumed to continue; if progress slows, customers could spend less or switch to rivals and open-source alternatives. Meanwhile, the enormous computing costs and long-term spending commitments could leave Anthropic exposed if future revenue fails to keep pace.
Explained simply
Anthropic is an artificial intelligence company that develops Claude, a family of AI models used by businesses and individuals. Training and running these models requires large amounts of computing power, which can be costly.
The company sells access to its models and related services to businesses, while competing with other AI developers and freely available open-source models. Its business prospects depend in part on whether it can keep attracting customers and improving its models as the technology and competition change.
AI Business Companies Technology