advertisement
OpenAI Signs 457MW Deal With AMD-Powered Cloud Provider TensorWave
TensorWave, a cloud provider based in Las Vegas, has signed a six-year agreement to build 457 megawatts of artificial intelligence computing capacity for OpenAI across several data centre sites in the United States. The infrastructure will run entirely on AMD’s new Instinct MI455X graphics processors. The announcement came on Monday, a day before Nvidia was due to release its quarterly financial results.
TensorWave has taken a different route from much of the AI cloud industry. While most providers have built their infrastructure around Nvidia chips, the company chose to work exclusively with processors from its rival, AMD. That decision has now put it in a position to supply computing capacity to one of the world’s biggest AI developers.
Under the agreement, all 457 megawatts will be reserved for OpenAI. This means OpenAI will have dedicated access to the infrastructure rather than sharing it with other customers. The systems will be built using AMD’s Helios rack-scale technology. Neither company has disclosed the value of the agreement or where the data centres will be located.
advertisement
The scale is worth putting into perspective. Airtel Africa’s data centre under construction at Tatu City, outside Nairobi, is expected to be the largest in East Africa, with a planned capacity of 44 megawatts. TensorWave’s commitment to OpenAI is more than ten times that size.
For TensorWave, this is its largest deployment so far. President and co-founder Piotr Tomasik called the 457-megawatt agreement a validation that AMD infrastructure can meet the scale and reliability demands of major AI companies.
The deal also places TensorWave within a much larger arrangement: the six-gigawatt strategic partnership between OpenAI and AMD. The company says it is the first publicly announced neocloud provider to support that partnership. Neoclouds are specialist cloud providers that focus on supplying the computing power needed to develop and run AI systems.
advertisement
For OpenAI, the agreement adds to its access to AMD-powered computing infrastructure as demand for AI services grows. “As demand for AI continues to grow, expanding access to reliable, large-scale compute is critical,” said Uday Ruddarraju, OpenAI’s chief technology officer for compute.
AMD also has a financial stake in TensorWave’s growth. Its venture arm, AMD Ventures, co-led the company’s $350 million Series B funding round in June alongside investment firm Magnetar. The round valued TensorWave at $1.55 billion.
The company had already established a sizeable AMD-based operation before landing the OpenAI agreement. Its existing infrastructure includes a training cluster with 8,192 Instinct MI325X processors. TensorWave also says it has secured more than two gigawatts of long-term data centre capacity. Among its customers are Fireworks AI and Luma AI.
advertisement
Vamsi Boppana, AMD’s senior vice-president for AI, described TensorWave as “a strong partner in scaling AMD Instinct infrastructure.” The deal comes as AMD looks to gain ground in a market where Nvidia has long been the dominant supplier of AI chips.
The timing puts that competition into focus. Nvidia is scheduled to report its fiscal third-quarter financial results on Tuesday, making the TensorWave announcement another development to watch as the competition for AI computing infrastructure intensifies.
The first computing capacity under the agreement is expected to come online in the first half of 2027. The full deployment is scheduled for completion by early 2028.