HomeFull speed ahead. Keep track of advancements in every frontier
CodexOpenAI—StargateOpenAI—OpenAI—GPT modelsOpenAI—OpenAI—gpt-ossOpenAI—CodexOpenAI—GPT modelsOpenAI—
Stargate
Courtesy of OpenAI · Press kit, editorial use (opens openai.com)

Stargate

by OpenAIUS

DeployStage 4 of 5

The Abilene, Texas campus is running and growing. More US and overseas sites are being built.

Updated 30 Oct 2025Checked 25 Sep0 updates this week

Milestones

Next · Full ~10 GW US buildout
  1. Stargate announced at the White House ($500B, ~10 GW)21 Jan 2025Complete.
  2. Stargate UAE (1 GW campus in Abu Dhabi) announcedMay 2025Complete.
  3. First Abilene buildings running NVIDIA GB200 racksSep 2025Complete.
  4. Five additional US sites announced (~7 GW planned total)23 Sep 2025Complete.
  5. Full ~10 GW US buildoutTarget 2029Not yet reached.

Most important updates

  • 30 Oct 2025
  • 23 Sep 2025
  • 22 Jul 2025
  • 22 May 2025
  • 21 Jan 2025

Upcoming

  1. 2029Full ~10 GW US buildout (next)

Current obstacles

  • Power and grid hookupsGigawatt campuses need new power plants and power lines. Waiting for grid hookups and turbines sets the pace.
  • Financing at huge scaleHundreds of billions in spending relies on loans and partners, betting on future AI revenue.

Physics limits

  • Almost all power becomes heatNearly every watt a chip draws ends up as heat. A 72-GPU rack draws about 120 to 130 kW, over 10x an old server rack, so air can't cool it; liquid loops must.
  • Light is too slow for far-apart GPUsSignals in fibre travel about 200 km per millisecond. Training syncs GPUs many times a second, so one tightly coupled cluster must sit on one site or a few nearby ones.
  • A gigawatt is a power plant1 GW running all year is about 8.8 TWh, the output of a large nuclear reactor. Sites need dedicated generation or new high-voltage lines, which take years to build.

How it works

4 parts
Server racks and network cabling inside a Stargate data hall (OpenAI photo)
Server racks and network cabling inside a Stargate data hall (OpenAI photo)Courtesy of OpenAI · Press kit, editorial use (opens openai.com)
Campus

Buildings full of GPU racks

Each site packs hundreds of thousands of NVIDIA GPUs into racks of 72 that act as one computer, joined by fast networks into one cluster.

Cooling

Liquid, not air

Racks draw over 100 kW each, so coolant is piped straight to the chips in closed loops that are filled once instead of evaporating water.

Power

Gigawatts per site

A 1 GW site uses what a large nuclear plant produces. Abilene pairs a grid connection with on-site gas generation.

Money

Partners share the bill

OpenAI, SoftBank, Oracle and MGX fund it. Oracle and SoftBank build and run most sites, and OpenAI leases the capacity.

Update log

6 updates

Thu 30 Oct 2025

  • Minor: BlogHardware

Tue 23 Sep 2025

  • Major: BlogHardware

Thu 31 Jul 2025

  • Minor: BlogPartnership

Tue 22 Jul 2025

  • Major: BlogPartnership

Thu 22 May 2025

  • Major: BlogPartnership

Tue 21 Jan 2025

  • Major: PressPartnership

About OpenAI

The team behind Stargate

OpenAI

ChatGPT, GPT models, Codex, Stargate

OpenAI makes ChatGPT and the GPT models behind it. ChatGPT's launch in November 2022 made AI chatbots mainstream. It also builds huge data centers through the Stargate project.

  • Makes ChatGPT and the GPT-5.x models. GPT-5.6 came out in July 2026.
  • Raised $122B at an $852B valuation in March 2026. Amazon, NVIDIA and SoftBank were the biggest investors.
  • Lining up tens of gigawatts of computing power through Stargate and chip deals with NVIDIA, AMD and others.
Founded
201511 yrs
Headquarters
United States
Status
Private
Valuation
$852BprivateMar 2026
Raised
$179B4 rounds
Last round
Venture round · $122BMar 2026$852B post
Works in
AIFrontier modelsComputing
Coverage
4 programs · 24 updateslatest 9 Jul 2026checked 25 Sep
Lead investors
AmazonNVIDIASoftBankThrive CapitalMicrosoft
Partners
Broadcom
Investors
MicrosoftSoftBankAmazon
People
Sam AltmanCEO and co-founderGreg BrockmanPresident and co-founderJakub PachockiChief Scientist
Primer