Building the compute infrastructure for the Intelligence Age

Source: OpenAI Blog

Stargate Overview

Stargate is OpenAI’s long‑term effort to build the compute foundation required to deliver the benefits of AGI broadly and reliably to the world. To meet the accelerating demand for AI across consumers, businesses, developers, and governments, we are expanding our compute footprint and bringing new capacity online faster. We are building together with partners, local communities, and the broader infrastructure ecosystem to stay ahead of shortages for the emerging compute‑powered economy.

Milestones

• January 2025: Announced a commitment to secure 10 GW of AI infrastructure in the United States by 2029.
• One year later: Surpassed that milestone, adding >3 GW in the last 90 days alone as demand for AI continues to accelerate.

“The only responsible way to meet this demand is to build more compute, faster.”

Our strategy remains unchanged because the reality has not changed: bringing the benefits of AI to more people requires more compute.

Why Compute Matters

• Core input that makes advanced AI possible.
• Enables:
  • Training better models
  • Reliable serving
  • Performance improvements
  • Cost reductions over time
  • Delivery of more powerful tools to more people

It sits at the center of our AI flywheel:

More compute → Better models → More usage → Better products & revenue → Reinvest in infrastructure

Financing models and partnership structures may evolve, but what matters is capacity coming online at scale, on time, and with flexibility as technology and demand evolve.

Partner‑Centric Approach

No single company can build the infrastructure for the Intelligence Age alone. Building AGI infrastructure at this scale requires coordination across the entire ecosystem:

• Local communities & utilities
• Energy providers & chipmakers
• Cloud providers & “neoclouds”
• Construction firms & investors
• Skilled trades & public‑sector partners

Our approach is intentionally partner‑centric. We want the ecosystem to succeed alongside us because delivering the compute required for the Intelligence Age will take all of us working together.

What Partnerships Deliver

• Faster progress
• More efficient builds
• Reduced execution risk
• Preserved optionality as technology changes

As we look beyond the initial 10 GW goal, OpenAI and our partners are evaluating potential data‑center locations across the country and will significantly expand our compute capacity in the years ahead.

Community Impact

Local Upside

AI infrastructure should create clear local benefits:

• Good jobs
• Strong schools
• New local revenue
• Responsible energy planning
• Careful water stewardship
• Early engagement with state and local leaders

Recent Community Investment

• Program launch: Donation to the Port Washington‑Saukville Education Foundation in Wisconsin (partnered with Vantage Data Centers and Oracle).
• Purpose: Support student opportunity, educator resources, and workforce‑readiness programming.

This is the first of many local investments we expect to make in communities connected to Stargate projects. We aim for long‑term, practical partnerships shaped by community priorities.

Workforce Development

Our work with North America’s Building Trades Unions (NABTU) and affiliated skilled‑trades unions reflects the same approach. Building AI infrastructure at this scale requires a trained workforce that can deliver safely, quickly, and to a high standard. These projects can create new job pathways into the Intelligence Age for skilled workers and apprentices—not just engineers and researchers.

The Abilene, Texas Site

Our Stargate site in Abilene, Texas illustrates how frontier AI infrastructure can be built quickly, responsibly, and in partnership with the surrounding community. It also shows that growth and stewardship can go together.

Water Stewardship

• Uses closed‑loop cooling (sealed pipes, recirculated water) instead of traditional evaporative cooling towers.
• One‑time initial fill per building ≈ two Olympic‑sized swimming pools.
• Annual water use at full build‑out ≈ a medium‑sized office building or four average households.

Responsible infrastructure matters because compute directly shapes what AI systems can do.

Translating Compute into Smarter Models

Our latest and smartest model yet, GPT‑5.5, was trained at the flagship Stargate site in Abilene, Texas.

• Infrastructure: Operates on Oracle Cloud Infrastructure and runs NVIDIA GB200 systems.
• Impact: Brings intelligence closer to people worldwide and to businesses of all sizes, enabling them to “just build things.”

GPT‑5.5 helps close the capability overhang—the productivity gap between power users of AI and those who are not—ensuring the economic opportunities of AI are broad‑based.

The countries and companies that can build and scale compute will help shape the future of AI.

Looking Ahead

• More compute → Stronger models
• Better systems → Wider adoption

We remain committed to expanding our compute capacity responsibly, in partnership with communities, and with a focus on creating lasting local benefits.

## Our Mission

We serve those models faster and more reliably. Greater efficiency helps us lower the cost of delivering intelligence over time. Together, those improvements make it possible for more people and businesses to use AI to build, learn, discover, and solve problems.

## The Stargate Initiative

That is the point of **Stargate**: to build the infrastructure needed to bring the benefits of AI to everyone.

## Progress & Outlook

- We are ahead of schedule, but the work is just beginning.  
- Demand is growing.  
- The opportunity is enormous.  

The communities that help build this infrastructure should share in the upside.

## Building the Intelligence Age

The Intelligence Age will not arrive on its own. It has to be built. And we are building it alongside our partners and local communities for the benefit of all of humanity.