The Economist warns Anthropic won't post huge profits any time soon: growth costs more than it earns

By The Economist · October 26, 2026.

The only material available is the headline of The Economist's article: "Don't expect supersize profits from Anthropic any time soon". We have not been able to read the text, so we do not know what figures, margins or arguments the publication puts forward. For that reason, we neither attribute nor invent them.

The only verifiable element is the angle: according to The Economist, very large profits at Anthropic are not imminent. This is the publication's thesis, not an accounting figure we can confirm here.

Our reading, clearly as opinion: the thesis is plausible and consistent with how the frontier-model business works. In general, leading labs reinvest a large share of revenue in compute, training and infrastructure. In addition, with reasoning models the inference bill grows, because many intermediate thinking tokens have to be paid for. Fast-rising revenue does not mean profits will rise as fast.

What this does not mean also matters. The absence of outsized profits in the short term is not a sign of failure. It is the investment phase of a technology whose value, in our view, will show over the long term. That includes faster science, better health and greater abundance, while the economic and labor transition remains hard. The underlying question is who will capture the intermediation margin once the cost of reasoning falls.

Once the full text is accessible, its figures should be checked before drawing firmer conclusions. With only the headline, we prefer a short, honest piece to filler.