OpenAI and Nvidia have announced plans to build the world’s most powerful AI supercomputer — a groundbreaking project with 10 gigawatts of capacity and $100 billion in investment, redefining the scale of artificial intelligence.
The Largest AI Infrastructure Project Ever
At the heart of this initiative is Nvidia’s next-generation Vera Rubin platform, which will power the new “AI factory.” With a scale comparable to multiple nuclear power plants, the system will provide unprecedented computational resources to train and run advanced AI models.
For years, AI progress has been limited not by ideas, but by a lack of computing power. This supercomputer aims to end that bottleneck, enabling the development of models of unmatched size, speed, and complexity.
A Bold Step Forward
“Everything starts with compute,” said Sam Altman, CEO of OpenAI, underscoring his vision that raw processing power is the catalyst for AI breakthroughs.
Jensen Huang, CEO of Nvidia, called the project the “necessary foundation for the next stage of AI development,” highlighting its role in accelerating global technological progress.
Transformative Potential
The first phase of the supercomputer will go live in late 2026, providing researchers with an unprecedented sandbox for experimentation and innovation.
With 10 gigawatts of dedicated capacity, this infrastructure could unlock advances that once belonged to science fiction — from next-generation healthcare and energy solutions to breakthroughs in personalized learning and scientific discovery.
