Lambda, Microsoft seal multibillion‑dollar AI deal powered by NVIDIA chips
Lambda has said it has agreed a multibillion‑dollar, multi‑year pact with Microsoft to deploy tens of thousands of Nvidia GPUs, including GB300 NVL72 systems, expanding Azure AI capacity amid a broader industry race for high‑end accelerators.
Lambda has announced a multibillion‑dollar, multi‑year agreement with Microsoft to deploy AI infrastructure powered by “tens of thousands” of Nvidia GPUs, including the new GB300 NVL72 systems, deepening the pair’s long‑running collaboration on high‑performance compute for artificial intelligence.
In a statement, the San Jose‑based cloud provider said it will deliver mission‑critical AI compute at scale to Microsoft, with deployments centred on Nvidia’s latest data‑centre platforms. The companies positioned the agreement as an expansion of their work together over more than eight years.
As per reports, some of the systems to be rolled out include Nvidia GB300 NVL72, devices designed for training and serving large‑scale models.
Why it matters for AI capacity
The deal has arrived amid an industry‑wide scramble for compute, with Microsoft separately having signed a $9.7bn, five‑year contract with data‑centre operator IREN to secure access to Nvidia’s latest chips.
Microsoft is also working on large Azure deployments based on GB300 NVL72 for advanced AI workloads.
About Lambda
Founded in 2012 by Stephen Balaban, Lambda has grown from selling on‑premises GPU servers to operating a cloud used by AI developers to train, fine‑tune and run models.
The company raised a $480M Series D this year with participation from Nvidia and other investors, and has previously secured a $500M GPU‑backed loan to expand its fleet.
Lambda’s relationship with Nvidia extends beyond equity: last month, reports indicated a $1.5B arrangement under which Nvidia has leased back tens of thousands of GPUs from Lambda over four years, reflecting intense demand for top‑tier accelerators.