The neocloud FUD is over and Jensen Huang just made it very clear that neoclouds are central to how Nvidia wants AI infr…
Milk Road AI Twitter · Milk Road AI (@MilkRoadAI) · 2026-07-02
Milk Road AI argues that Nvidia's new revenue-sharing and credit-support partnerships with neoclouds including CoreWeave, Nebius, and Sharon AI validate the neocloud business model and are a structurally bullish signal for the sector.
Appears in
Extraction
Topics: neocloud-investmentnvidia-strategyai-infrastructurecompute-market
Claims
- Nvidia is partnering with neoclouds to deploy multi-tenant AI factories through revenue-sharing and credit-support arrangements, lowering the capital barrier for neocloud operators.
- Jensen Huang framed AI infrastructure as shifting from one-off model training to always-on token production, fundamentally changing datacenter economics toward continuous 'AI factory' models.
- Revenue per gigawatt, not raw GPU ownership, is the key differentiating metric for neocloud operators, with Anthropic-deployed capacity generating more revenue per gigawatt than OpenAI-deployed capacity.
- Nvidia sharing in neocloud revenue upside validates the economics because Nvidia is now betting its own recurring revenue on the ability of these clouds to sell tokens profitably.
- Sharon AI is deploying up to 40,000 Nvidia GB300 GPUs under a six-year collaboration where Nvidia and lenders fund much of the hardware in exchange for a share of cloud revenue.
Key quotes
Nvidia is partnering with AI clouds, the neoclouds to deploy large scale, multi‑tenant AI factories.
it proves the economics are real, because Nvidia is now sharing in the upside instead of just selling hardware and walking away, it is literally betting its own recurring revenue on these clouds being able to sell tokens profitably.
The real metric is revenue per gigawatt, not just how many GPUs you own.