NVIDIA has introduced a new revenue-sharing and credit-support business model with AI cloud providers aimed at accelerating global access to large-scale AI infrastructure as demand for enterprise AI, inference computing, and AI-native applications continues to surge worldwide.
The new model is designed to help AI cloud providers deploy NVIDIA-powered AI infrastructure faster without the lengthy timelines and capital constraints traditionally associated with hyperscale data centre development and GPU infrastructure expansion.
Under the arrangement, participating AI cloud providers procure NVIDIA infrastructure and offer NVIDIA-powered cloud services to customers, while NVIDIA earns revenue through both hardware sales and a share of the cloud revenue generated from supported AI capacity.
The initiative reflects the rapidly evolving economics of the AI industry, where demand is increasingly shifting from large-scale AI model training toward continuous production inference workloads. Industry analysts note that inference infrastructure is becoming one of the most critical segments within the global AI ecosystem as enterprises deploy AI agents, generative AI systems, autonomous workflows, and real-time AI applications at scale.
NVIDIA stated that the new framework is intended to support startups, AI model developers, enterprises, sovereign AI initiatives, and regional AI cloud providers seeking faster access to enterprise-grade AI computing infrastructure.
Among the first companies adopting the model are Sharon AI, Inc. and FIRMUS. Sharon AI plans to deploy up to 40,000 NVIDIA Grace Blackwell GB300 GPUs as part of its AI cloud infrastructure expansion strategy. Meanwhile, FIRMUS is developing a large-scale DSX AI factory campus in Batam, Indonesia, expected to scale to approximately 360 megawatts of capacity and support up to 170,000 NVIDIA GPUs over time.
Industry observers believe the announcement highlights the growing transition toward what NVIDIA increasingly describes as “AI factories” — large-scale AI infrastructure environments designed to operate continuously and generate AI inference tokens at industrial scale. Unlike traditional cloud infrastructure optimized primarily for storage and compute, AI factories are being architected specifically for persistent AI workloads, autonomous systems, and large-scale inference operations.
The move also underscores intensifying global competition around AI infrastructure deployment as governments, enterprises, hyperscalers, and AI-native startups race to secure GPU capacity, AI cloud infrastructure, and sovereign AI computing capabilities.
NVIDIA continues to strengthen its position at the centre of the global AI infrastructure ecosystem through its GPU platforms, AI computing systems, networking technologies, software frameworks, and enterprise AI infrastructure solutions. The company’s Grace Blackwell architecture is expected to play a major role in powering next-generation AI workloads, autonomous AI systems, generative AI platforms, and large-scale enterprise AI deployment globally.
