OpenAI was founded in 2015 partly as a counterweight to the concentration of AI research inside companies such as Google. A decade later, that rivalry has given way to a more pragmatic arrangement: Google will distribute OpenAI’s frontier models through Google Cloud. ¹

The partnership builds on an earlier, quieter link. OpenAI said last year that it was already using Google’s cloud to run ChatGPT and its API, even as Microsoft remained its primary cloud partner. Then, less than six months ago, Microsoft’s exclusive license to OpenAI intellectual property ended; Amazon moved quickly to host OpenAI models, putting the company on a path toward all three major cloud platforms. ¹

Now Google is rolling out GPT-6 and GPT-6.1 models — Astra, Sol and Luna — in private preview for selected enterprise customers, with broader availability in its Model Garden expected in coming weeks. The models will run entirely on Google infrastructure and be purchased through the Cloud Marketplace, a useful shortcut for companies already committed to Google Cloud spending. ²

For OpenAI, the deal is a distribution play: access to developers and corporate buyers who may prefer Google’s cloud environment. For Google, it is a hedge against customers who want the best available model for a task rather than a single-vendor AI commitment. Google says the addition makes it the only hyperscale cloud provider offering all three leading closed-source large language models alongside more than 200 other options. Sundar Pichai (sundarpichai)³

The agreement also reaches beyond cloud procurement. OpenAI models are headed to Antigravity, Google’s agent-powered coding tool, where Gemini and Anthropic’s Claude already operate. That puts Google in the unusual position of hosting its own flagship alongside two of its fiercest AI competitors — and making customer choice the product.