

OpenAI has inked a deal with Oracle Corp. to develop 4.5 gigawatts’ worth of data center capacity in the U.S.
The companies announced the agreement today. It’s part of Stargate, an initiative to build a network of stateside artificial intelligence data centers for OpenAI. The project will see the ChatGPT developer and its partners invest up to $100 billion in 10 gigawatts of AI infrastructure.
In a blog post, OpenAI stated that it now expects to “exceed our initial commitment thanks to strong momentum with partners.” But though the company’s long-term goals for the project may have grown more ambitious, it’s reportedly scaling back its near-term plans. The Wall Street Journal reported on Monday that the scope of the project’s initial phase has been “sharply” reduced.
According to the paper’s sources, OpenAI has set a goal of building a small data center by year’s end. It’s believed the facility will likely be located in Ohio. OpenAI is currently working on two additional data center campuses in Texas.
The AI provider’s construction push is being complicated by disagreements with SoftBank Group Corp., the Journal cited its sources as saying. The Japanese conglomerate became OpenAI’s biggest backer this year after it committed to leading a funding round worth up to $40 billion. SoftBank owns the Stargate trademark.
According to the Journal, several key terms of the partnership have not yet been finalized. That includes the role of SB Energy, a SoftBank subsidiary, in the construction initiative. OpenAI and SoftBank have reportedly not yet agreed on how many Stargate data centers will be built on sites tied to SB Energy.
The ChatGPT developer stated in today’s blog post that “with SoftBank, we’re moving quickly on site assessments and reimagining how data centers are designed to power advanced AI.”
So far, OpenAI has reportedly raised a quarter of the $40 billion that its SoftBank-led funding round could be worth. The AI provider is expected to launch an effort to raise the remaining $30 billion on July 28. SoftBank will provide up to three quarters of the sum.
Oracle, for its part, is reportedly poised to generate more than $30 billion annually from its AI infrastructure deal with OpenAI. It expects to reach that revenue milestone within three years.
In addition to the 4.5 gigawatts of capacity announced today, Oracle is responsible for building one of OpenAI’s Texas data center campuses. The site is expected to house up to 400,000 of Nvidia Corp.’s GB200 Grace Blackwell accelerators. Each chip features two Blackwell B200 graphics cards and a 72-core Grace central processing unit.
OpenAI disclosed today that Oracle has started moving GB200-equipped server racks into the data center. Furthermore, the ChatGPT developer’s researchers are already using the hardware to run AI training and inference workloads. The data center is expected to become fully operational in mid-2026.
Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE’s Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities.
Founded by tech visionaries John Furrier and Dave Vellante, SiliconANGLE Media has built a dynamic ecosystem of industry-leading digital media brands that reach 15+ million elite tech professionals. Our new proprietary theCUBE AI Video Cloud is breaking ground in audience interaction, leveraging theCUBEai.com neural network to help technology companies make data-driven decisions and stay at the forefront of industry conversations.