OECD warns AI markets risk entrenched Big Tech power

The OECD’s new brief on AI markets is effectively a scoreboard for the industrial side of the AGI race. It stitches together evidence that, despite rapid innovation and a flood of new models, the most strategically important layers of the stack—frontier foundation models, cutting-edge accelerators, and the data and cloud platforms they run on—remain dominated by a small set of firms. That concentration isn’t just a business story; it shapes who can afford to train near-frontier systems, who controls safety practices, and which jurisdictions have real leverage over the trajectory of intelligent systems.

In the near term, this kind of analysis will inform competition regulators deciding whether to treat AI as “just another software market” or as critical infrastructure closer to energy or telecoms. Remedies could range from enforcing interoperability and data portability, to scrutinizing exclusive compute and model-access deals between hyperscalers and labs. For incumbents racing toward AGI, the risk is that aggressive vertical integration—owning chips, models, cloud, and distribution—starts to look like monopolization on paper, inviting structural remedies. For challengers, especially open-weight and regional players, the brief is intellectual ammunition to argue for policies that lower capital and compute barriers to entry. How these competition debates land will influence whether the next wave of frontier systems emerges from a few mega-labs or from a more plural ecosystem of labs, clouds, and open consortiums.