NVIDIA has launched a revenue-sharing financing model for AI cloud providers. This marks a shift in how large-scale artificial intelligence infrastructure is funded and deployed. The structure is designed to help cloud operators access NVIDIA’s computing systems while linking payments to future revenue generated from AI services.
The model arrives at a time when demand for AI computing is moving beyond training large models toward continuous inference workloads. These workloads require constant access to high-performance systems to generate outputs at scale for enterprises and AI-focused companies.
Under the new framework, AI cloud providers can acquire NVIDIA hardware with credit support. Repayment is connected to revenue earned from cloud services built on that infrastructure.
NVIDIA continues to receive standard revenue from hardware sales. It also gains a share of the income generated through the supported cloud capacity. The company positions this structure as a way to ease upfront financial pressure on operators building large AI systems.
The approach targets a long-standing challenge in the AI sector. Many startups and emerging cloud providers struggle to finance large data centre deployments that require heavy capital investment before generating returns.
The financing model reflects a broader transition in the AI industry. Companies are moving from experimental model development toward production systems that run continuously.
These systems support tasks such as inference, fine-tuning, and real-time AI services. The shift has increased demand for scalable computing infrastructure that can be deployed quickly and expanded as usage grows.
NVIDIA highlighted that AI-native firms and cloud providers now require flexible access to compute rather than traditional long-term infrastructure commitments.
Two companies have been named among the early collaborators under this model. Sharon AI plans to deploy up to 40,000 NVIDIA Grace Blackwell GB300 GPUs. The company aims to build a sovereign large-scale AI computing infrastructure using NVIDIA’s systems.
Firmus is developing a DSX AI factory campus in Batam, Indonesia. The facility is expected to scale up to 360 megawatts and support as many as 170,000 NVIDIA GPUs. The project is designed to serve the growing global demand for AI computing capacity.
These developments highlight the scale of infrastructure NVIDIA is targeting through its financing strategy. The model links deployment to future usage rather than relying only on upfront capital commitments.
NVIDIA also pointed to AI-native companies such as Baseten, Fireworks AI, and Together AI as examples of growing demand for scalable compute infrastructure. These firms depend on continuous access to accelerated computing as customer workloads expand.
Sharon AI co-founder and CEO James Manning described the partnership as a key step in building large-scale compute infrastructure. Firmus co-CEO Tim Rosenfield noted that AI companies require scalable and efficient compute systems to remain globally competitive.
The financing approach signals a broader shift in how AI infrastructure is built and funded. It reflects growing demand for flexible models that support the rapid expansion of cloud-based AI services.
Also Read: OpenAI Introduces Jalapeño; Signals New Challenge to NVIDIA’s AI Dominance with its AI Chip