NVIDIA unlocks GPU compute at scale through a revenue-sharing model – lowering the barrier to enterprise AI infrastructure
Category: AI Infrastructure
What it is
NVIDIA is enabling AI cloud providers to deploy its infrastructure through a
revenue-sharing model – no large upfront hardware spend required. Sharon AI is
deploying 40,000 Grace Blackwell GB300 GPUs under this structure. Firmus is
building a 360-megawatt AI factory campus in Batam, Indonesia scaling to
170,000 NVIDIA GPUs. The model is designed to close the gap between surging AI
demand and constrained compute supply.
Why it Matters for Enterprises
GPU scarcity has capped enterprise AI ambitions in 2026. This model accelerates
new supply - the compute crunch rationing mid-size enterprise access should
ease in late 2026. Enterprises planning OCI GenAI or Oracle AI deployments
should revisit availability timelines in Q3.
