OpenAI Integrates Advanced Models and Codex into Google Cloud

On Wednesday, October 21, OpenAI announced the integration of its frontier artificial intelligence models and Codex coding agent into Google Cloud's Model Garden. The deployment brings GPT-6 Astra, GPT-6.1 Sol, and GPT-6 Luna to select enterprise customers in a private preview, with general availability scheduled for the coming weeks across the platform.[1][3]

The agreement places OpenAI's technology on all three of the largest United States cloud infrastructure providers, marking a major milestone in the company's enterprise strategy. Businesses can now purchase OpenAI's models directly through the Google Cloud Marketplace, paying for the services using their existing Google Cloud spending commitments without negotiating new contracts.[1][3][4]

All usage will appear on a single Google Cloud invoice, streamlining procurement and billing for enterprise technology departments. Customer prompts and outputs remain secured within the client's established data perimeter, and Google has explicitly confirmed it will not use this sensitive enterprise data to train its own competing artificial intelligence models.[1][5]

Expanding enterprise distribution

The Google Cloud integration represents a significant shift in OpenAI's distribution strategy as it actively reduces its historical reliance on Microsoft Azure. While OpenAI emphasized in its announcement that Microsoft remains its primary cloud partner, the company is aggressively targeting corporate customers across competing infrastructures to maximize its market share.[1][5]

"This partnership is an important step in making our most capable models and Codex available where customers are already building," said Greg Brockman, OpenAI's president and co-founder. The move acknowledges that enterprise clients prefer to access artificial intelligence tools within the environments they already operate.[1]

The models will run on dedicated Google Cloud hardware, ensuring high performance and integration with Google's built-in enterprise data governance frameworks. Developers can already connect Codex to Google's platform, and the models are slated for integration into Google Antigravity and the Gemini Agent.[3][5]

The expansion follows a similar move in April, when OpenAI brought its models to Amazon Web Services' Bedrock platform. That AWS integration occurred shortly after a renegotiated agreement with Microsoft allowed OpenAI to cap revenue-share obligations and offer services through competing platforms.[3][4]

Competing on the same shelf

The partnership creates an unusual dynamic where Google and OpenAI, direct competitors in the generative artificial intelligence market, are collaborating closely on distribution. Google's Gemini models actively challenge OpenAI's products across consumer applications, software development tools, and high-margin enterprise services, making the shared hosting arrangement a delicate balancing act.[1][5]

"We are focused on giving customers a broad selection of models, including those from OpenAI, so they can choose the right model for any given job," stated Thomas Kurian, chief executive of Google Cloud. Google now positions itself as the only hyperscale cloud provider offering all three leading closed models, as its platform already hosts Anthropic's Claude alongside its native Gemini.[1][4]

For enterprise buyers, the arrangement eliminates the need to migrate data or establish new vendor relationships to access OpenAI's technology. By offering competing systems side-by-side, Google Cloud aims to capture the infrastructure revenue regardless of which artificial intelligence model a client selects.[1][2]

The integration specifically targets software development lifecycles and workplace productivity. By bringing Codex agentic workflows natively to Google Cloud, OpenAI is positioning its tools to automate complex coding tasks and engineer autonomous software agents at scale.[2][3]

The shifting cloud landscape

The cloud infrastructure market is increasingly defined by artificial intelligence hosting capabilities, with providers competing fiercely to offer the most comprehensive model libraries. Amazon Web Services previously expanded its OpenAI offerings following a massive $38 billion commitment between the two companies that was disclosed to the public in November 2025.[3][4]

While neither Amazon nor OpenAI has released public revenue figures generated by their arrangement, the rapid expansion to Google Cloud suggests strong enterprise demand for multi-cloud model access. Corporate technology officers have consistently requested the ability to deploy OpenAI models without being locked into a single cloud vendor.[3][5]

The Google Cloud deal also highlights the massive compute requirements necessary to run frontier models at an enterprise scale. By distributing the workload across Azure, AWS, and Google Cloud, OpenAI can better manage the infrastructure demands of its growing corporate client base.[2][4]

OpenAI's own operations have previously utilized Google's infrastructure. Last year, the company disclosed that ChatGPT and its application programming interface were running on Google's cloud infrastructure, handling millions of daily queries and laying the technical groundwork for this formal commercial partnership.[3]

What happens next

As the private preview concludes over the next 14 to 21 days, Google Cloud will open the OpenAI models to its broader customer base. Enterprise technology teams will then be able to evaluate GPT-6 Astra and Codex directly against Gemini 4 Argon and Claude 3.5 within the same testing environment.[1][3]

The success of the partnership will likely be measured by how many existing Google Cloud customers allocate their committed spending to OpenAI's models rather than Google's native offerings. This dynamic will test whether infrastructure providers can sustainably profit from hosting their direct competitors while maintaining their own proprietary systems.[1][4]

The financial mechanics of the deal remain undisclosed, but industry analysts note that such integrations typically involve complex revenue-sharing agreements. Google Cloud secures the compute and hosting revenue, while OpenAI captures the model licensing fees, creating a mutually beneficial arrangement despite their broader rivalry over the $250 billion generative artificial intelligence market.[2][5]

For the broader software industry, the availability of Codex agentic workflows on Google Cloud lowers the barrier to entry for automated development. Engineering teams can now build sophisticated, autonomous software agents using OpenAI's reasoning capabilities without leaving their established Google Cloud security perimeters.[2][3]

The integration signals that the artificial intelligence industry is rapidly moving toward a model-agnostic cloud ecosystem. The deciding factor for enterprise adoption will increasingly be the specific performance, latency, and cost-efficiency of the models themselves, rather than the underlying cloud platform that hosts them.[1][4]