Google Cloud Adds OpenAI Models to Enterprise Platform

Google Cloud has officially broadened its model offerings by incorporating OpenAI systems onto its infrastructure alongside its proprietary Gemini models and Anthropic's Claude. This expansion positions Google as the only major cloud vendor providing access to all three leading AI model families in one ecosystem. Beginning today, select enterprise clients can register for a private preview to access three newly available OpenAI options: GPT-6 Astra, built for demanding workloads like advanced research and coding; GPT-6.1 Sol, a lower-cost variant engineered for similar benchmark performance; and GPT-6 Luna, a rapid, economical model tailored for general everyday tasks.

The integration introduces key administrative and technical benefits for software engineers and enterprise architects. Beyond standard API access, the newly added OpenAI models are set to appear across Gemini for Business and Google Antigravity. To streamline operations for enterprise customers, all consumption charges will consolidate onto a single unified Google Cloud bill, directly counting toward existing enterprise contract commitments. Additionally, Google clarified that any business data routed through these services will remain strictly private and will not be harvested for training future models.

While this strategy boosts Google Cloud's market positioning, it underscores an internal conflict over internal hardware allocation. Splitting valuable compute infrastructure to support top market competitors creates friction with Google's native Gemini engineering units. This allocation of server capacity toward rivals highlights the tension between cloud unit revenue and internal AI development, a conflict that reportedly played a role in driving two-billion-dollar researcher Noam Shazeer back to OpenAI.