NVIDIA Opens Compute Access via Revenue-Share Model
NVIDIA is introducing a revenue-sharing partnership model that allows AI cloud providers to procure its infrastructure and resell services to startups, enterprises, and research organizations. The model addresses capital constraints that have historically limited emerging AI companies' access to large-scale compute. Early partners Sharon AI and Firmus are deploying tens of thousands of NVIDIA GPUs through this arrangement.
TL;DR
- NVIDIA is launching a new business model pairing revenue-sharing and credit support to unlock compute access for AI startups and enterprises
- Sharon AI is deploying up to 40,000 NVIDIA Grace Blackwell GB300 GPUs under the partnership
- Firmus is building a DSX AI factory campus in Batam, Indonesia, expected to scale to 360 megawatts and up to 170,000 NVIDIA GPUs
- The model targets AI-native companies needing immediate access to accelerated computing for training, fine-tuning, and inference workloads
Why It Matters
As AI workloads shift from model development to production inference at scale, compute demand is accelerating. Emerging AI companies have historically struggled to secure capital for infrastructure investment, creating a bottleneck. NVIDIA's partnership model removes that friction by enabling cloud providers to finance and operate large compute facilities, making accelerated infrastructure accessible to a broader ecosystem.
Business Impact
The revenue-sharing structure gives NVIDIA both direct product revenue and a recurring, usage-linked earnings stream tied to cloud service adoption. For AI companies, the model eliminates delays from site selection, power procurement, and hardware deployment, enabling faster time to production. Cloud providers gain a differentiated service offering and capital support to build regional AI infrastructure.
Key Implications
- NVIDIA is shifting from a pure hardware vendor model to a platform-and-revenue-share model, creating recurring revenue streams tied to AI workload utilization
- Regional AI infrastructure buildout will accelerate as capital barriers lower, potentially fragmenting compute access across multiple cloud providers rather than concentrating it with hyperscalers
- AI-native companies and inference providers gain faster access to production-scale compute, reducing time-to-market for AI applications and services
What to Watch
Monitor how many additional cloud providers adopt this partnership model and what scale they reach. Track whether this model successfully captures demand from AI-native companies or whether hyperscalers' existing compute offerings remain dominant. Watch for regional variations in deployment, particularly in geographies like Southeast Asia where Firmus is building.
Subscribe to the newsletter
The latest stories and analysis, delivered to your inbox.
Free. No spam. Unsubscribe any time.
