“`html
OpenAI Secures Compute Capacity via Google Cloud and CoreWeave Deal
In a strategic move to bolster its artificial intelligence capabilities, OpenAI is leveraging a deal involving Google Cloud and CoreWeave to secure additional computing capacity.CoreWeave will provide computing resources to Google’s cloud unit, which will then make that capacity available to OpenAI, according to Reuters.
Strategic Partnership Details
this arrangement follows reports of a capacity agreement between Google Cloud and OpenAI, despite the competitive landscape with Google’s DeepMind AI unit and the challenge ChatGPT poses to Google’s search business. Google will directly provide compute capacity to OpenAI through its own cloud platform, in addition to the resources obtained via CoreWeave.
Did You Know? openai’s demand for computing power is so immense that it’s driving meaningful growth in the data center industry, with providers like Oracle scrambling to secure more capacity.
The structure of this deal, were CoreWeave leases to Google, who then releases to OpenAI, mirrors previous arrangements involving OpenAI, Oracle, and Microsoft. In June 2024, OpenAI was leasing servers from Microsoft, which in turn leased them from Oracle.At that time, OpenAI had an agreement with Microsoft to exclusively use its cloud resources for training large language models (LLMs).
Diversifying Cloud partnerships
OpenAI has as diversified its cloud partnerships. The decision to utilize compute from CoreWeave, despite not having an exclusive agreement with Google Cloud, is highly likely driven by an urgent need for increased capacity. Sources indicate that OpenAI had previously sought a direct deal with Google Cloud but was hindered by its existing commitment to Microsoft.
This diversification strategy is also evident in OpenAI’s other initiatives, such as the $500 billion Stargate project, announced earlier this year in collaboration with SoftBank, Oracle, and Abu dhabi’s MGX. The first Stargate campus will be located in Abilene, Texas, with Oracle reportedly planning to purchase approximately 400,000 Nvidia GB200s for the site.
Global Expansion of AI Infrastructure
Beyond the United States, OpenAI is exploring the progress of Stargate campuses in other locations, including the UAE. OpenAI’s chief strategy officer is currently touring the Asia Pacific region, including India, Japan, South Korea, Australia, and Singapore, to assess potential data center investment incentives. In Europe, OpenAI is reportedly in discussions with the UK, France, and germany regarding Stargate expansion after initial success in the US.
Pro Tip: Keep an eye on the evolving landscape of AI infrastructure partnerships, as they often signal shifts in technology strategies and market dynamics.
OpenAI also signed a $4 billion cloud deal with CoreWeave in May 2025, building upon an earlier $1.9 billion deal with the neocloud provider.
Oracle’s Data Center Expansion
Financial services firm TD Cowen reported this week that Oracle is seeking approximately 5GW of data center capacity in the US for OpenAI, separate from the Stargate project. Oracle CEO Safra Catz confirmed this during a recent earnings call, stating that OpenAI is contributing to Oracle’s significant cloud growth, even with Stargate still “in formation.”
| Partner |
|---|
