AI supercomputer · serves as: Compute, AI Inference, AI acceleration.
Which group Stargate belongs to and how it is built
An AI compute cluster / data center is a hardware subcategory covering large-scale computing infrastructure built from thousands of graphics processing units (GPUs) and AI accelerators, high-bandwidth networking, power delivery and cooling systems. Such installations are used to train and run (infer) large artificial-intelligence models; their scale is measured in megawatts or gigawatts of electrical power and in compute throughput.
An AI supercomputer is a component type covering an integrated, large-scale computing system built from clusters of graphics processors and dedicated AI accelerators connected by high-speed networks. It is designed for parallel training and running of large AI models (e.g. language and multimodal models), emphasising throughput, energy efficiency and scalability.
target by 2029
10 data-center buildings
Stargate (The Stargate Project) is a program to build AI computing infrastructure in the United States, announced on 21 January 2025. The planned investment reaches up to $500 billion by 2029, with an initial commitment on the order of $100 billion.
The equity funders are OpenAI, SoftBank Group, Oracle and the investment firm MGX. SoftBank holds financial responsibility while OpenAI holds operational responsibility; Masayoshi Son serves as chairman. OpenAI and SoftBank each took roughly 40% ownership ($19 billion each), while Oracle and MGX contributed $7 billion each.
The technology partners are Arm, Microsoft, NVIDIA, Oracle and OpenAI. The first campus is being built in Abilene, Texas (ultimately ten data-center buildings), with planned expansion to additional US states and countries.
The project targets up to 10 GW of compute capacity. The infrastructure is intended to train and run OpenAI’s large AI models.
