1 Terawatt Annual Compute Target

The facility targets 1 terawatt of AI compute per year—roughly 50x current global output—requiring 100,000–1 million wafer starts per month and producing 100–200 billion chips annually.