Nvidia says OpenAI will deploy its AI infrastructure at massive scale, reaching 12GW by 2030

Nvidia CEO Jensen Huang said OpenAI has committed to deploying its AI infrastructure at scale by 2030, with current and planned deployments totaling about 12GW of Nvidia compute. If the partnership expands, this could grow to 16GW, representing a business opportunity of around $600 billion by 2030.

Nvidia announced a partnership with SB Energy to secure land, power, and building shell (LPS) capacity at the PORTS-Pike tech park in Portsmouth, Ohio, USA, for its compute, with OpenAI as the tenant. The initial phase will provide 4.25GW of capacity, with each generation of systems using about 1.5 million Nvidia GPUs and generating $150-200 billion in revenue.

Nvidia emphasized that building AI factories requires a complete set of resources, including chips, packaging, memory, networking, and land, power, and building shells. The company is applying its approach to securing semiconductor resources to LPS, ensuring long-term capacity. The park can support multiple system upgrades over a 20-year cycle, and the partnership may expand to 8GW in the future.