OpenAI's GPT-6 Models Enter Private Preview on Google Cloud, Completing Its Spread Across the Three Largest US Clouds

What this means

The mechanism is distribution through cloud commitments. Enterprise AI budgets increasingly sit inside committed cloud spend, and a model that cannot draw on that money loses deals it would otherwise win. OpenAI gains reach into accounts that run on Google. Google gives up some Gemini exclusivity in exchange for keeping workloads and margin on its own infrastructure. The loser on exclusivity is Microsoft, whose Azure advantage on OpenAI models now rests only on timing and integration depth. Investors already expected multi-cloud distribution for frontier models, because Anthropic followed the same path. The new part is that Google put a direct rival's models inside its own agent products. That suggests Google sees model choice, not Gemini lock-in, as what keeps enterprise customers. The consensus reads this as a win for OpenAI. The cloud that controls routing may hold the stronger long-run position, since it collects revenue regardless of which model wins a given task. The likelier outcome is that within months the hyperscalers compete on routing, evaluation and governance tooling layered over interchangeable frontier models, and that per-token prices on the cloud marketplaces move toward OpenAI's direct pricing. That depends on general availability arriving on schedule. If Google limits OpenAI models to a narrow preview through the first quarter of 2027, the read is wrong.