TensorWave to provide OpenAI with six years of AMD AI capacity

TensorWave agreed to provide OpenAI with 457 megawatts of dedicated computing capacity for six years across multiple U.S. data-center sites. The infrastructure will use AMD Instinct MI455X GPUs and Helios rack systems, with initial capacity expected in the first half of 2027 and full deployment targeted for early 2028. The companies did not disclose the contract’s value.

The agreement creates a dedicated deployment channel for AMD GPUs and expands TensorWave’s role in OpenAI’s AMD infrastructure buildout. For AMD, the deal provides a disclosed multi-year demand commitment for its AI hardware, although no contract revenue figure was provided.

  • 1The agreement covers 457 megawatts of computing capacity for six years.
  • 2The deployment will use AMD Instinct MI455X GPUs across multiple U.S. data-center sites.
  • 3The first capacity is expected online in the first half of 2027, with the full rollout targeted for early 2028.
  • 4OpenAI will receive dedicated capacity reserved for its workloads rather than shared cloud infrastructure.
  • 5TensorWave describes the project as its largest deployment to date.

Sources