Nvidia (NVDA) is taking a bold step by allowing AI startups to utilize its powerful chips while compensating the company with a share of future revenues. This significant initiative was outlined in a recent blog post published on July 1, which positions Nvidia as a key investor in the AI expansion rather than a typical hardware seller.
Transitioning from Sales to Revenue Sharing
Traditionally, Nvidia generates income through a one-time transaction when selling graphics processing units (GPUs). However, this new program introduces a recurring revenue model that supplements their conventional sales. Cloud partners acquire Nvidia systems and provide access to startups that lack the funds to establish their own data centers. In this setup, Nvidia takes a percentage of the revenue produced by these cloud services powered by their chips.
“This structure … provides NVIDIA with a recurring, usage-linked earnings stream,” stated an excerpt from the blog co-authored with Chief Financial Officer Colette Kress.
Building Strategic Partnerships
This initiative leverages Nvidia’s latest AI compute model and diversifies its clientele, especially in light of recent decreases in orders from larger buyers. Startups that participate in this program will remain interconnected with Nvidia's technology for the long term. For instance, Sharon AI is preparing to install as many as 40,000 Grace Blackwell GB300 chips under this agreement.
Investors Leopold and Situational Awareness recently acquired a 19.9% stake in Sharon AI, and shortly thereafter, both entities announced a strategic partnership with Nvidia. Additionally, Firmus is currently constructing a substantial 360-megawatt facility in Batam, Indonesia, designed to host up to 170,000 GPUs, operational by the first quarter of 2027.
Assessing Competitive Dynamics and Market Risks
This revenue-sharing approach provides Nvidia with a competitive edge, especially as rivals enter the field. Notably, China has trained a large AI model without relying on Nvidia chips, prompting buyers to explore alternative, cost-effective solutions. Such dynamics raise questions about the sustainability of Nvidia's model amidst growing competition.
Nvidia's financial commitment to the AI sector mirrors what critics refer to as circular financing, having pledged as much as $100 billion to OpenAI. Furthermore, the company holds approximately 7% of CoreWeave, a prominent customer purchasing their chips. Analysts have likened this strategy to the vendor financing seen during the dot-com boom, which experienced criticism from skeptics regarding the formation of an AI bubble.
The numbers involved are substantial. Morgan Stanley projects that AI capital investments from major tech companies will exceed $800 billion by 2026 and could reach $1.1 trillion by 2027, challenging even the U.S. defense budget.
Despite these ambitious projections, market reactions remained stable with NVDA closing at $194.69 on July 2, the last trading day before the holiday break, maintaining a market cap of around $4.8 trillion, slightly below its 52-week peak. Observers will be keen to see how this innovative revenue-sharing program impacts Nvidia’s bottom line in the upcoming quarters.
