We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
OpenAI Blog·Oct 30·5 min read
Expanding Stargate to Michigan
TL;DR
The United States faced a capacity gap in AI computing infrastructure needed to support large-scale model training and deployment. OpenAI is building a one-gigawatt campus in Michigan—a facility consuming one billion watts of power—to expand Stargate, its distributed computing infrastructure.
✦ Why It Matters
Engineers can plan AI workloads knowing expanded domestic computing capacity will support larger-scale model training and inference operations.
Key Takeaways
- Plan large-scale AI projects with confidence in expanded domestic computing resources.
- Monitor Stargate capacity announcements to optimize infrastructure allocation decisions.
- Consider regional AI infrastructure availability when designing distributed training pipelines.
- Track power and cooling requirements for future data center planning.
Full Summary
OpenAI announced an expansion of Stargate, a large-scale computing infrastructure project designed to provide the computational resources required for training and running advanced AI models. The new Michigan campus will operate at one gigawatt of power capacity, making it a major data center facility.
This expansion addresses a critical bottleneck: the shortage of specialized hardware and power infrastructure needed to support increasingly large AI workloads. The project strengthens domestic AI infrastructure by reducing dependence on international computing resources and creating a foundation for future AI research and deployment.
Beyond technical capacity, the initiative drives economic growth in the Midwest through job creation and capital investment in the region. The concrete outcome is a new physical facility that increases America's total AI computing capacity and establishes Michigan as a hub for AI infrastructure development.