Nvidia Corporation has announced a large-scale program that fundamentally changes the approach to monetizing computing power for artificial intelligence. The tech giant is now ready to provide its GPUs to young companies on terms of partial payment from future revenue. This is not just a new pricing plan—it is a strategic move that transforms the chip manufacturer from a component supplier into a full-fledged investor and partner.
From One-Time Sales to Recurring Revenue
Nvidia's traditional model was built on one-time sales of graphics processors. The new initiative creates a permanent source of income that will complement primary hardware sales. Large cloud providers purchase the equipment, lease computing power to startups, and Nvidia receives a fixed share of the profits generated by these processors from end customers. As noted by the company's Chief Financial Officer Colette Kress, this scheme ensures regular income directly dependent on the utilization rate of the equipment.
The program relies on an updated AI platform and becomes accessible to a wide range of companies. This is particularly relevant amid cost optimization by some major clients, who are increasingly seeking more flexible financing options.
Global Scale and Fierce Competition
The scale of the project is impressive. Sharon AI plans to install up to 40,000 Grace Blackwell GB300 chips under this program, while Firmus is building a 360-megawatt campus in Batam, Indonesia, to house up to 170,000 graphics processors. Such projects are already being called "sovereign computing" for markets outside the United States.
However, locking clients into its own ecosystem is becoming critically important amid growing competition. Chinese developers recently successfully trained a large-scale neural network without using American processors, and many global clients are actively testing alternative solutions. Nvidia is essentially betting on long-term startup lock-in to its software and chips.
Technological Boom or Financial Bubble?
Such financial schemes evoke associations with cross-financing mechanisms among independent experts. It is known that Nvidia has committed to investing up to $100 billion in the OpenAI project and holds an approximately 7% stake in CoreWeave, a major buyer of its equipment. Michael Burry and other skeptics believe that this structure only fuels concerns about a bubble in the AI sector.
According to Morgan Stanley estimates, the total spending of the world's largest IT corporations on artificial intelligence technology development will exceed $800 billion by 2026, and could reach a record $1.1 trillion by 2027—a sum comparable to the U.S. defense budget. Meanwhile, the stock market showed relative stability: Nvidia shares closed at $194.69, and the company's market capitalization approached $4.8 trillion, though this is still slightly below the historical highs of the current year.
My expert opinion: This model is a brilliant move by Nvidia, allowing it not only to smooth out sales cyclicality but also to create a closed loop of dependency for startups. However, in the long term, the program's success will depend on whether young companies can generate enough profit to pay back the giant. If the AI bubble bursts, Nvidia risks becoming the largest creditor in a bankrupt ecosystem.