NVIDIA is rolling out a new revenue-sharing and credit-support model with AI cloud partners to accelerate the deployment of large-scale, multi-tenant AI factories. The structure allows partners to build and operate DSX AI factories while NVIDIA earns standard product revenue plus a share of the cloud revenue generated on supported capacity.
Revenue-Sharing Model Details
The model aligns economics between NVIDIA and AI clouds serving AI-native enterprises, ISVs, and research organizations. It provides NVIDIA with a recurring, usage-linked earnings stream tied to actual compute consumption. For customers, it delivers faster access to full-stack accelerated computing by bypassing traditional delays in site selection, power procurement, construction, and hardware deployment. The factories are designed for continuous operation, generating tokens at scale for production inference, model training, post-training, fine-tuning, and high-volume agentic workloads.
First Partner Deployments
Sharon AI is the initial partner deploying under the framework, scaling to as many as 40,000 NVIDIA Grace Blackwell GB300 GPUs. Firmus is constructing a major DSX AI factory campus in Batam, Indonesia, with plans to reach 360 megawatts of power capacity and support up to 170,000 NVIDIA GPUs. These commitments illustrate how the model enables rapid hyperscale buildouts through aligned capital and infrastructure partners.
Broader Ecosystem Acceleration
NVIDIA is inviting additional capital and infrastructure partners to power similar AI factories that can serve customers across regions. High-growth AI natives such as Baseten, Fireworks AI, and Together AI exemplify the types of organizations that gain immediate production-grade capacity without upfront capital barriers. The approach expands platform adoption while creating shared upside in the growing AI cloud services market.
Data Points for Investors
This new model establishes a dual revenue path for NVIDIA — infrastructure sales combined with ongoing participation in partner cloud revenue — supporting scalable AI compute access globally. Early metrics include Sharon AI’s 40,000 GB300 GPU deployment and Firmus’ 360 MW / 170,000 GPU campus target.
High-net-worth investors focused on NVIDIA’s ecosystem strategy, recurring revenue potential, and AI infrastructure expansion now have concrete partner commitments and a clear framework for how capital partners are being mobilized to build multi-tenant AI factories at scale.
Additional information is available at blogs.nvidia.com.