TensorWave will supply OpenAI with 457MW of AMD-powered AI capacity
TensorWave, an AMD-backed AI cloud provider, will supply OpenAI with 457MW of computing capacity using AMD's Instinct MI455X GPUs. The six-year deal, part of an existing partnership, will be deployed across US data centers, with the first capacity online by mid-2027. TensorWave, which uses only AMD hardware, raised $350M in its Series B funding round.
How this was made

The 30-second read
Why it matters
The deal is a multi-year, single-tenant compute commitment for OpenAI using AMD Instinct MI455X GPUs, with staged deployment from H1 2027 through early 2028.
Market read
Traders may reassess AI accelerator demand durability and AMD’s role in frontier-AI compute buildouts as a named, large-capacity deployment is scheduled to ramp in 2027-2028.
What to watch
Execution risk (rack-scale Helios integration, power availability, and deployment schedule) could delay utilization, muting the near-term earnings impact despite the headline MW figure.
Background
TensorWave, an AMD-backed AI cloud provider, is the first publicly named neocloud provider under OpenAI’s existing AMD partnership.
Ticker impact
OpenAI capacity for 457MW will run on AMD Instinct MI455X GPUs under an AMD partnership, with deployment timeline through early 2028.
Shares may see supportive sentiment as the market prices in incremental MI455X/Instinct adoption tied to OpenAI’s compute buildout.
The article discloses a large, multi-year GPU capacity commitment tied to AMD hardware, but it does not provide contract value or site-level ramp details that would quantify revenue impact.
Market effects
Reinforces the trend of large, rack-scale GPU deployments for frontier AI and may support sentiment across AI infrastructure supply chains.
US data-center buildout demand could benefit US power, networking, and colocation ecosystems tied to GPU capacity expansions.
Signals continued global competition for frontier-AI compute capacity, potentially affecting GPU procurement expectations beyond the US.
Counterpoint
Without contract value or named sites, the market may treat this as a capacity-planning announcement rather than a near-term revenue catalyst for AMD.
Key entities
- companyTensorWave
AI cloud provider designing AMD Helios rack-based data-center sites and supplying OpenAI with 457MW of compute capacity.
- companyOpenAI
Frontier AI customer receiving dedicated GPU capacity under its existing AMD partnership.
- companyAMD
Supplies Instinct MI455X GPUs powering the 457MW compute capacity for OpenAI via TensorWave.





