NVIDIA Launches AI Compute Partnership: From Chip Seller to Compute Bank
On July 1, The Information reported that NVIDIA is leveraging its strong balance sheet to provide financial guarantees for emerging GPU cloud providers, transforming from a chip vendor into the 'central bank' of the AI compute ecosystem. NVIDIA officially announced a new business model combining revenue sharing and credit support, allowing AI cloud providers to procure NVIDIA infrastructure without fully bearing upfront capital expenditures, and to offer compute services to AI-native companies, model developers, and enterprise clients.
Revenue Sharing and Financial Guarantee Mechanism
According to the report, NVIDIA promises to repurchase unsold GPU capacity at agreed prices if cloud providers cannot find enough AI developers to rent computing power. In exchange, NVIDIA takes a percentage of the cloud providers' revenue, with the share declining over time. Internally, the program is called the 'AI Compute Partnership.' NVIDIA confirmed the program's existence, positioning it as the 'DSX AI Factory' model for cross-regional, high-utilization, multi-tenant accelerated computing.
Analysts note that this move kills two birds with one stone: NVIDIA's backstop effectively acts as a credit enhancement tool, enabling emerging cloud providers that previously struggled to secure bank loans for GPU financing to raise larger capital for data center construction. Since GPUs are typically the most expensive component of AI data centers, low-credit buyers face significant loan hurdles. NVIDIA's guarantee solves both GPU and data center financing issues simultaneously.
Initial Participants and Massive Deployments
Sharon AI and Firmus are the first participants in this model. Sharon AI plans to deploy up to 40,000 NVIDIA Grace Blackwell GB300 GPUs. Firmus is building a DSX AI factory campus in Batam, Indonesia, expected to expand to 360 MW and host up to 170,000 NVIDIA GPUs. These deployments directly demonstrate NVIDIA's latest progress in converting compute demand into deployable, financeable infrastructure.
NVIDIA stated that emerging AI companies have historically faced severe restrictions in accessing capital-intensive infrastructure—even long-term commitment contracts often fail to unlock financing for compute procurement. The new model promises to realign economic structures, allowing AI-native companies, model developers, and inference providers to quickly access full-stack accelerated computing without waiting for traditional infrastructure construction cycles.
Strategic Background: Reducing Hyperscaler Dependency, Binding Emerging Cloud Providers
Currently, a handful of hyperscalers—Amazon, Microsoft, SpaceX, Oracle, Meta, and Google—purchase the majority of NVIDIA's chip production. However, several of these companies are developing competitive in-house AI chips, posing a potential threat to NVIDIA. To reduce dependency, NVIDIA has spent years cultivating emerging GPU cloud providers like CoreWeave. The AI Compute Partnership is a continuation and deepening of this strategy.
According to previous reports from The Information, NVIDIA has also been negotiating financial guarantees for OpenAI's potential large data center lease in Ohio, which could cost up to $500 billion if fully built. NVIDIA has already invested billions in equity stakes in emerging cloud providers, with total deals involving billions of dollars for capacity repurchase with CoreWeave and Lambda. In September 2024, NVIDIA committed to buying all of CoreWeave's unsold capacity through 2032 for $6.3 billion if no tenants were found—a move that boosted CoreWeave's stock nearly 30% within a week. According to NVIDIA's May regulatory filing (covering the quarter through April), the company added another $3.5 billion to guarantee customer data center leases in exchange for stock purchase rights. Overall, NVIDIA is building multi-layered interest-binding mechanisms: equity investments, capacity repurchases, lease guarantees, and now revenue sharing—each layer deepening financial ties with downstream cloud providers, enabling NVIDIA to directly capture incremental revenue from AI compute commercialization beyond chip sales.
