OpenAI Models Are Now On All Three Major US Cloud Platforms

OpenAI's rollout of its models to Google Cloud's Model Garden and Antigravity coding tool, alongside existing Azure and AWS offerings, marks the first time its AI services are available on all three major U.S. cloud platforms after the Microsoft‑exclusive license lapsed earlier this year. For website owners, this multi‑cloud availability means AI‑driven traffic can originate from a broader set of IP ranges and service endpoints. Bots powered by OpenAI’s GPT‑4o or Claude‑compatible models may now be instantiated on Google Cloud’s infrastructure, adding Google‑owned CIDR blocks to the mix of Azure and AWS addresses that already generate substantial API calls, content generation, and automated form submissions. Consequently, traditional bot‑filtering rules that focus only on Azure or AWS ranges will miss a growing slice of legitimate and malicious OpenAI traffic. **Actionable tips:** 1. Update your llms.txt to include the new Google Cloud model garden domains (e.g., `modelgarden.googleapis.com`) and Antigravity endpoints, explicitly allowing crawlers you trust while blocking unknown OpenAI‑generated agents. 2. Enrich your bot‑tracking firewall with Google Cloud’s latest IP ranges (downloadable from Google’s public JSON list) and tag any requests bearing OpenAI user‑agent strings for separate logging. 3. Deploy a lightweight verification step—such as a challenge‑response token—on high‑value forms to differentiate human visitors from AI‑generated submissions originating from any of the three cloud providers.