Nvidia CEO Jensen Huang calls Elon Musk the 'ultimate GPU' (www.businessinsider.com)

Nvidia CEO Jensen Huang praised Elon Musk on the BG2 podcast, calling him the “ultimate GPU” for his ability to hold the many interdependencies of building massive AI infrastructure in his head — from financing and land to power and procurement. Huang was referring to Musk’s xAI Colossus II project near Memphis, which Musk bills as the “world’s first Gigawatt AI training cluster.” Colossus II reportedly already contains at least 200,000 Nvidia GPUs and aims to scale toward a million GPUs, a scale Huang said could reach a gigawatt before others do. The remarks matter because they highlight two technical and industry realities: first, the extraordinary systems-level complexity of assembling hyperscale AI training clusters (power provisioning, facility shells, financing, hardware supply); and second, Nvidia’s central role as the primary GPU supplier for Big Tech’s AI race. A million‑GPU, gigawatt-class datacenter would dramatically increase training throughput and energy demand, intensifying competition for chips, racks, and power capacity and accelerating compute-driven model scale. Huang’s praise also underscores how individual leadership and integrated decision-making can shorten timelines for such projects — with broad implications for supply chains, energy infrastructure, and who leads next-generation foundation-model development.