80% of Investment Went to OpenAI and Anthropic: The Market Looks Broad, but the Center of the New Economy Is Already Very Narrow

The decision in “The Optical Illusion of the AI Market” depends on one criterion: hundreds of startups appear every week, but most of the money concentrates in OpenAI and Anthropic, because without foundation intelligence the rest of the ecosystem simply cannot exist.

The “Why it's not just a ranking but a map of the AI market's next phase” scene leads to a working conclusion: it is dangerous to look only at the giants: the next wave may grow in applied markets where the model is not yet the product — in companies with their own data and a customer process.

The practical meaning of “OpenAI and Anthropic as the centre of the new AI economy” is that investors finance models and compute because the ecosystem needs foundation intelligence, but that centrality does not mean the giants will capture all the value in applied markets.

For the “What will happen to the programming market” scene, the decisive point is this: Codex and Claude Code take over more and more work, while small tools risk becoming a feature inside a platform; the developer does not disappear, but the market pays less for manual boilerplate and more for architecture, data, and responsibility.

For the “What will happen to applied professions” scene, the decisive point is this: a lawyer, physician, salesperson, or analyst gets an assistant that speeds up individual operations, but full replacement is harder because of trust, regulation, and ambiguous decisions, and this is where new companies connect a model to industry logic.

In the context of “Physical AI: robots, cars, devices,” this criterion applies: physical AI requires manufacturing, sensors, and service, so owners of the foundation model cannot automatically capture all the value in that market.

The discussion of “Creative AI: video, audio, images” yields a practical test: creative AI depends on rights, brands, and distribution, so a single foundation model does not capture the market's value — it stays with those who control rights and the path to an audience.

The decision in “Early Forbes AI companies: who could lead the next wave?” depends on one criterion: the conclusion for a profession is the same. Do not compete with a model at the operation it performs best. Occupy the place where the result has to enter a real market and someone has to answer for it.