OpenAI, Oracle, and SoftBank accelerate Stargate in the US

  • Five new AI data centers in the US are powering Stargate, with three projects from Oracle and two from SoftBank.
  • Planned capacity is close to 7 GW and investment exceeds 400.000 billion euros, with the goal of 10 GW and 500.000 billion euros.
  • SoftBank plans up to 1,5 GW in 18 months; a possible 600 MW expansion in Abilene; and more than 25.000 direct jobs.
  • Selected after reviewing 300 proposals in 30 states; deployed with OCI and NVIDIA GB200 racks in Abilene.

Stargate AI data centers in the United States

The consortium of OpenAI, Oracle and SoftBank is hitting the gas in the United States with a coordinated deployment of new data centers under the Stargate initiative. The move reinforces the commitment to increasing the computing power that next-generation AI models demand.

With this announcement, the Stargate roadmap is significantly expanded: the planned capacity is located close to 7 gigawatts and the committed investment exceeds 400.000 million, directing the project towards its objective of 10 GW and 500.000 billion.

OpenAI, Oracle, and SoftBank plan five new AI data centers for the Stargate project.
Related article:
OpenAI, Oracle, and SoftBank power Stargate with five US data centers.

New locations and scope

Locations of the new Stargate data centers

Oracle will lead three strategic locations: the Shackelford County, Texas, Doña Ana County, New Mexico and a location yet to be announced in the Midwest. In addition, a 600 MW expansion next to the main campus in Abilene, Texas, to strengthen the anchoring of capacity in the region.

SoftBank, for its part, will develop two additional sites: Lordstown, Ohio and the Milam County, TexasThe Ohio project is already underway with a view to entering into operation as of 2026, while the one in Texas will rely on the quick build model of SB Energy, the energy subsidiary of the Japanese group.

Together, the new Oracle-powered sites can contribute more than 5,5 GW, and SoftBank's two are designed to scale up to 1,5 GW in about 18 months. It is a significant expansion that seeks to balance the speed of deployment, cost efficiency and scalability.

The three companies emphasize that these sites represent an important step to accelerate the deployment of AI infrastructure, with a pragmatic focus on timing of construction, power provisioning, and productionization of training and inference services.

Capacity, investment and employment

Stargate Program Capacity and Investment

With the five new centers, the initiative is very close to the planned 7 GW and exceeds the medium-term investmentThe 10 GW and 500.000 billion euro target is closer at hand, with a portfolio of additional locations still under evaluation.

The selection process began in January and has been competitive: applicants were reviewed more than 300 proposals in thirty states, prioritizing energy availability, connectivity, land, and construction deadlines. Companies are advancing that there will be more ads as new phases are finalized.

The deployment provides for the creation of more than 25.000 direct jobs in construction and operation, as well as tens of thousands of indirect jobs in the supply chain. For local communities, this translates into sustained industrial activity and demand for technical talent.

In Abilene, Texas, Oracle has already begun installing NVIDIA GB200 racks within Oracle Cloud Infrastructure (OCI), allowing for early training and inference loads to be launched and preparing for capacity increases as new batches of hardware arrive.

Company leaders have emphasized the approach: Sam Altman summarizes that AI will only fulfill its promise with the computing power adequate; Clay Magouyrk highlights that OCI's infrastructure is helping OpenAI to scale quickly; and Masayoshi Son puts emphasis on center design and the energy experience from SoftBank to support expansion.

In parallel, SoftBank articulates the energy leg with SB Energy to accelerate the availability of electrical power and storage, a critical aspect for operate on a large scale without compromising network stability or commissioning times.

The package of measures — new sites, expansions, chip supply and cloud agreements— puts Stargate in a strong position to deliver high-performance computing in the short and medium term, with the capacity to absorb new loads as the models evolve.

Everything points to the collaboration between expanding in phases will continue to expand in phases, with additional locations and technical improvements that maintain the balance between scalability, cost and speed of deployment, key elements for the next wave of AI applications.

The Stargate expansion crystallizes a large-scale industrial strategy: five new centers, capacity close to 7 GW, investments exceeding 400.000 billion and a construction schedule that is already underway in Ohio and Texas. With OCI's technological base, the NVIDIA GB200 racks and the energy contribution of SB Energy, the project accelerates to convert that ambition of 10 GW and en tangible infrastructure y employment in the United States.


Follow us on Google News