Nvidia expands from chips to AI infrastructure

Nvidia has moved beyond the role of a standalone GPU vendor by selling a broader AI infrastructure platform that includes the Vera CPU, Rubin GPU, NVLink, InfiniBand, Ethernet networking, systems, algorithms and CUDA software. Management said fiscal 2028 revenue should grow approximately 70% and described demand as supply-constrained. Revenue opportunity per gigawatt has expanded from roughly $18 billion per gigawatt with Hopper to $25 billion per gigawatt with Blackwell to $40 billion per gigawatt with Vera Rubin.

The company also disclosed partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to mobilize over $500 billion in third-party capital for AI infrastructure. A SoftBank Energy campus in Ohio will host Nvidia compute for OpenAI under 20-year leases, while hyperscalers are described as carrying cloud backlogs greater than $2 trillion.

Broadcom, AMD, Marvell and Intel are each pursuing different parts of the same AI buildout. Broadcom reported AI semiconductor revenue of $16.70 billion, up 221%, and focuses on custom XPUs and networking. AMD is pushing EPYC CPUs, Instinct GPUs and Helios racks, while Marvell supplies custom silicon and optical interconnect. Intel is relying on Xeon demand and its foundry turnaround, though Intel Foundry posted a $2.1 billion quarterly operating loss.