OpenAI has secured a major computing supply agreement with TensorWave, an AI cloud provider backed by AMD's venture arm. The partnership will deliver 457 megawatts of processing capacity built on AMD's Instinct MI455X GPUs, distributed across multiple US data centre locations. The contract extends for six years, though neither party disclosed financial terms.

This arrangement represents an expansion of OpenAI's existing partnership with AMD, which had previously announced a broader collaboration. TensorWave becomes the first publicly named neocloud provider—a cloud company built specifically around GPUs for artificial intelligence—operating under that umbrella agreement. The first systems are scheduled to begin operations in the first half of 2027, with full deployment expected by early 2028.

Built for Helios

TensorWave is architecting the deployment around AMD's Helios racks, which consolidate multiple GPUs into unified rack-scale systems. These racks will be connected via Ethernet, and the entire capacity will be dedicated exclusively to OpenAI with no multi-tenant sharing.

We're excited to advance our previously announced partnership with AMD through this agreement with TensorWave

Uday Ruddarraju, OpenAI's chief technology officer for compute

Ruddarraju took on responsibility for OpenAI's data centre and compute infrastructure in October.

TensorWave's largest project to date

Based in Las Vegas, TensorWave operates exclusively with AMD hardware. The company was the first public cloud provider to deploy AMD's MI300X GPUs, and has since expanded its portfolio to include the MI325X and MI355X models. The OpenAI contract represents its most substantial deployment yet.

In June, TensorWave closed a $350 million Series B funding round backed by investors including Magnetar, AMD Ventures and Nexus Venture Partners. The company has announced it has locked in more than 2 gigawatts of long-term data centre capacity across its operations.

AI infrastructure is entering a new phase, where the ability to deliver hundreds of megawatts of advanced compute reliably and at speed will separate the companies that can support frontier AI from those that cannot

Piotr Tomasik, TensorWave's president and co-founder

Vamsi Boppana, AMD's senior vice president for AI, described the arrangement as a "landmark deployment" for OpenAI in the announcement.

Source: The Next Web