Nvidia is expanding beyond chip sales with a new program that gives emerging cloud providers financial backing to build AI data centers, while allowing the company to share in their future revenue.
The agreement, called the AI Compute Partnership, lets smaller cloud operators obtain Nvidia hardware without paying the full upfront cost. In return, Nvidia receives a revenue share that declines over time and agrees to repurchase unused GPU capacity at a fixed price if the providers cannot attract enough demand from AI developers and enterprise clients.
The structure gives Nvidia a larger role in financing AI infrastructure at a time when demand for high-performance computing remains strong but access to capital is uneven. It also marks a shift from one-time hardware sales toward recurring revenue tied to actual AI compute usage.
How the partnership works
Under the arrangement, cloud providers can use Nvidia’s financial support as credit assurance when raising money for data centers. The repurchase commitment effectively reduces lender risk because Nvidia is backing part of the unused capacity.
That is important because GPUs are typically the most expensive part of AI computing facilities. Smaller or lower-rated companies often struggle to finance large GPU purchases and long-term data center leases. Nvidia’s support can help these firms secure capital more easily and begin construction sooner.
For Nvidia, the benefit is broader than selling more chips. The company gains a share of cloud revenue and strengthens its position inside the AI compute supply chain, from hardware supply to infrastructure financing and usage-based economics.
Early participants
Two companies have joined the program as early participants.
Sharon AI plans to deploy about 40,000 Nvidia Grace Blackwell GB300 GPUs. Firmus is building a 360-megawatt AI factory in Batam, Indonesia, using about 170,000 units of Nvidia hardware.
These deployments would add significant centralized AI computing capacity to the market. They also show how Nvidia is using financing tools to help newer cloud companies compete in a sector dominated by large technology groups.
Why Nvidia is changing its model
The new program reduces Nvidia’s reliance on major customers such as Amazon and Microsoft, both of which are developing their own AI chips. By backing younger cloud providers, Nvidia can create a more diversified customer base and deepen dependence on its hardware ecosystem.
The move also gives Nvidia exposure to the downstream AI compute market. Instead of earning money only when chips are sold, the company can participate in the revenue generated when those chips are rented out to AI startups, model developers and corporate users.
The global AI chip market is projected to reach about $100 billion in 2026. The inference segment, which involves running AI models after they have been trained, is expected to account for the largest share, at roughly 58%. Nvidia’s revenue-sharing model places the company closer to that growing pool of usage-based spending.
A widening web of financial ties
The AI Compute Partnership builds on several earlier Nvidia financing arrangements.
The company previously made a $6.3 billion commitment to buy unused capacity from CoreWeave if it is not rented by 2032. Nvidia has also made equity investments in companies such as Lambda, while its researchers lease GPU servers from some of these cloud providers.
Regulatory filings for the quarter ending in April showed another $3.5 billion allocated to guarantee customer data center leases in exchange for stock purchase rights.
Together, these arrangements show Nvidia building a multi-layered financial role in the AI infrastructure market. Its links now include equity stakes, capacity repurchase commitments, lease guarantees and revenue-sharing agreements.
Impact on AI compute pricing
The program could increase the supply of high-performance computing capacity by helping smaller cloud providers finance large GPU deployments. If enough new capacity reaches the market, pricing pressure could emerge in some AI compute services.
That would matter for cloud companies, enterprise buyers and traders watching sectors tied to AI infrastructure. Lower computing costs could support more model development and deployment, but they could also challenge providers that rely on higher margins for GPU rental services.
The impact will depend on how quickly new facilities are completed, how much demand comes from AI developers, and whether enterprise adoption continues to absorb added capacity.
Relevance for decentralized compute projects
The development is also important for traders focused on digital assets linked to decentralized physical infrastructure and distributed compute networks.
A subsidized increase in centralized GPU supply could change the economics for decentralized alternatives. If centralized cloud capacity becomes more available or cheaper, distributed hardware networks may need to prove stronger advantages in cost, access, resilience or specialized use cases.
For now, Nvidia’s strategy shows that the company is not only defending its lead in AI chips. It is also helping finance the infrastructure that uses those chips, positioning itself as a central financial and technological force in the next phase of AI growth.
Explore how AI complements blockchain to understand evolving digital infrastructure beyond Nvidia’s expanding AI compute finance strategy.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
