Nvidia pauses AI cloud revenue-sharing deals within 2 months

Nvidia has paused its revenue-sharing deals with AI cloud companies less than two months after unveiling the program, a strategic reversal that could reshape how smaller cloud providers secure access to its scarce accelerators. The Santa Clara, California-based chipmaker announced the program in late June or early July, and the pause came Aug. 27, according to people familiar with the matter.

"Our entire supply chain is challenged... At this point we have supply for 70%. Our demand is much higher than that," Jensen Huang, Nvidia's chief executive, said on the company's earnings call this week.

The revenue-sharing model was designed to let AI cloud providers pay for Nvidia accelerators partly through future revenue, giving smaller players a path to capacity that hyperscalers like Amazon.com Inc., Microsoft Corp. and Alphabet Inc.'s Google secure through massive upfront purchases. Nvidia's decision to rethink the program comes as its supply commitments across its supply, infrastructure and partner network have climbed to $279 billion from $119 billion last quarter, with much of the increase tied to memory procurement, CFO Colette Kress said.

The pause lands days after Nvidia reported blockbuster fiscal second-quarter results. Revenue more than doubled to $96.22 billion, beating the $92.27 billion consensus, while data center revenue reached $89 billion, up 117 percent from a year earlier. The company's AI Clouds, industrial and enterprise, or ACIE, customers generated $40.3 billion in the quarter, up 138 percent, a sign that growth is broadening beyond the largest hyperscalers.

Why the pause matters

The revenue-sharing program was seen as a way for Nvidia to lock in demand from AI cloud startups and enterprises that cannot match hyperscaler purchasing power. Morgan Stanley analysts had identified the push into cloud revenue-sharing as a potential additional growth driver for Nvidia, calling the company's 70 percent fiscal 2028 growth forecast "remarkable" given it remains supply constrained.

Nvidia's decision to pause the program raises questions about the economics of the model, which ties Nvidia's revenue to the performance of its cloud customers. It also comes as the company faces questions about how it will allocate its constrained supply. Huang said demand is running ahead of what Nvidia can supply, making the 70 percent growth forecast effectively a supply-constrained number.

The program's future matters for the broader AI infrastructure market. AI cloud providers that were relying on revenue-sharing deals to secure Nvidia accelerators may now need to seek alternative financing or turn to competitors such as Advanced Micro Devices Inc. and in-house chips from hyperscalers. Nvidia's data center business, which accounts for about 92 percent of total sales, remains the engine of its growth, but the company is also expanding into server CPUs and robotics.

Nvidia shares rose 4.1 percent in after-hours trading following the earnings call, and the stock is up 12.4 percent so far this year. The company trades at about 17.9 times forward earnings, below AMD's 37.2 times and Intel's 46.2 times. At least 10 brokerages raised their price targets after the results, with Goldman Sachs lifting its 12-month target to $300 from $285 and Citigroup raising its target to $315 from $300.

The pause in revenue-sharing deals could weigh on AI cloud stocks that had counted on the program, while reinforcing Nvidia's control over its supply chain. Nvidia's gross margin, at 75 percent, is expected to ease to 74 percent this quarter and bottom out between 71 percent and 72 percent in the fourth quarter of fiscal 2027 as memory costs rise. The company has not yet disclosed a timeline for resuming or restructuring the program.

This article is for informational purposes only and does not constitute investment advice.