NVIDIA has unveiled a new business model designed to accelerate the deployment of large-scale, always-on AI compute infrastructure, addressing the capital and financing barriers that have historically constrained rapid AI production scaling. As artificial intelligence transitions from experimental development to continuous production inference, demand for token-scale computing has surged. This shift requires high-utilization, multi-tenant accelerated computing environments that can be deployed rapidly and operated efficiently. NVIDIA’s updated framework enables AI cloud providers to procure and deploy hardware directly for startups, enterprise clients, and independent software vendors. Under this revenue-sharing and credit-support structure, cloud partners sell AI-native services while NVIDIA collects standard hardware revenue alongside a recurring, usage-linked portion of cloud service earnings. This alignment accelerates platform adoption across the high-growth AI sector while providing NVIDIA with predictable, scalable revenue streams. The initiative is already materializing through dedicated DSX AI factory deployments. Sharon AI has partnered with NVIDIA to deploy up to 40,000 Grace Blackwell GB300 GPUs, aiming to deliver sovereign, large-scale compute for regional and enterprise workloads. Simultaneously, Firmus Technologies is constructing a 360-megawatt AI factory campus in Batam, Indonesia, designed to scale to 170,000 NVIDIA GPUs. These facilities will directly serve AI-native companies and inference providers that require immediate access to full-stack accelerated computing. By bypassing traditional infrastructure hurdles such as site selection, power procurement, and hardware commissioning, model builders, agent platforms, and enterprises can rapidly transition from pilot programs to production-scale deployment. Industry participants highlight the model’s capacity to support training, fine-tuning, and high-volume agentic inference. Companies such as Baseten, Fireworks AI, and Together AI exemplify the shifting demand toward elastic, production-ready compute. The revenue-sharing arrangement provides these emerging platforms with commercial flexibility and reliable infrastructure as they scale, while ensuring NVIDIA’s ecosystem expands in tandem with market adoption. Organizations seeking immediate capacity can coordinate directly with Sharon AI and Firmus to secure allocated resources. This partnership-driven infrastructure strategy positions NVIDIA to meet the escalating requirements of global AI development while establishing a sustainable, usage-based economic model for the broader compute ecosystem.