OpenAI distributes its models via Google Cloud

OpenAI will make its models and Codex available through Google Cloud, with purchases via Marketplace and use in enterprise environments. The agreement expands the publisher's distribution beyond Azure and targets enterprise AI workflows.

OpenAI will offer its AI models and its coding agent Codex via Google Cloud to reach enterprise customers already embedded in the Google ecosystem. Access will be through Model Garden and Google Cloud Marketplace, with general availability announced in the coming weeks. The partnership mainly targets companies wanting to consume these models within their existing cloud contracts without overhauling their entire infrastructure. For CIOs, it adds another option to an already crowded strategic choice set.

Un accord pensé pour les équipes déjà sur Google Cloud

The core of the announcement is simple: Google Cloud customers will be able to purchase and use OpenAI models through their usual environment, including leveraging their existing spending commitments. In practice, this reduces purchase friction and can accelerate adoption in large organizations, especially those managing strict procurement procedures.

Google Cloud specifies that these models will operate with its enterprise security and governance controls. Prompts, context, and responses will remain within the data perimeter defined by the client and will not be used to train foundation models, according to Google. Zero retention options are expected later, without a detailed timeline [to be verified].

Codex et les agents IA montent en gamme

The agreement is not only about text generation models. Codex, OpenAI's development agent, will also be integrated into software development workflows, opening the door to more operational uses: code assistance, automation of repetitive tasks, and integration into business processes. OpenAI and Google also mention use cases in regulated sectors, where governance remains decisive.

Greg Brockman, OpenAI's president and co-founder, praised a partnership that aims to put the most advanced models "where customers are already building." On his side, Thomas Kurian, head of Google Cloud, emphasizes the variety of models offered, "so they can choose the right model for any given job, helping to deliver optimal quality and lower costs." The idea is clear: let the customer choose the right engine for the task, rather than imposing a single provider.

La distribution d’OpenAI se démultiplie, avec Microsoft toujours en toile de fond

This announcement expands OpenAI's presence across multiple clouds, while Microsoft remains its main cloud partner. In April, the agreement between the two companies was modified to allow OpenAI to serve its products on other cloud providers. Microsoft still holds its position as the main partner, and OpenAI products continue to launch first on Azure, unless Microsoft cannot or does not want to support the necessary capabilities.

The move makes commercial sense. The more OpenAI multiplies its access points, the easier it becomes for companies to buy its models without changing their cloud stack for a single AI provider. At the same time, Google Cloud enriches its own AI platform with a player already attracting part of the market. The rapprochement thus has a very competitive flavor, even if it is presented as a gain of choice for customers.

Key points

  • General availability expected in the coming weeks.
  • Google Cloud Marketplace will serve as the purchase channel.
  • Existing spending commitments can be used.
  • Greg Brockman and Thomas Kurian advocate for more model choices.
  • OpenAI remains distributed beyond Azure since the April agreement.

By the numbers

  • 2 players — OpenAI and Google Cloud at the center of the announced agreement.
  • April — date of the agreement modification opening OpenAI to other clouds.
  • A few weeks — announced delay before general availability.

To read