Nvidia and OpenAI have formalized a landmark agreement to construct one of the most ambitious artificial intelligence infrastructure projects in history. A letter of intent signed September 22, 2025, outlines plans for OpenAI to deploy at least 10 gigawatts of Nvidia systems, comprising millions of GPUs designed to power next-generation AI models.
The initial infrastructure phase targets a launch in the second half of 2026 utilizing Nvidia’s Vera Rubin platform. This next-generation chip architecture is engineered specifically for massive-scale AI workloads. Nvidia has committed up to $100 billion to the initiative through a milestone-based investment structure. Capital will be released only as systems are deployed and operational benchmarks are achieved, rather than through upfront lump-sum payments.
This expansion carries significant implications beyond the technology sector. Facilities consuming 10 gigawatts or more will create substantial downstream effects across energy markets and data center real estate. For investors, the commitment represents both a major opportunity and concentration risk, with measurable results dependent on a timeline extending into late 2026.
Market participants are now monitoring three critical variables: whether the Vera Rubin platform delivers promised performance gains, if OpenAI’s compute demands maintain their current trajectory, and whether competitors can offer viable alternatives before the first phase becomes operational. The cryptocurrency sector remains unconnected to these specific developments, with no blockchain projects cited in the partnership terms.