Nvidia pauses revenue-sharing deals with AI cloud companies: Report

Nvidia has temporarily halted certain revenue-sharing agreements with AI cloud companies, according to a report by the Wall Street Journal. The chip manufacturer suspended the program last week, though it could potentially revamp it later or integrate it into another initiative. The new business model, which provides compute access to the rapidly expanding AI ecosystem, remains operational and continues to develop due to strong demand. Nvidia spokesperson confirmed the ongoing evolution of the program.

This decision follows less than two months since Nvidia's initial announcement of the financing initiative aimed at assisting small AI cloud firms with their financial requirements. The company sought to rent out compute capacity to cloud customers in case they couldn't sell it themselves, thereby ensuring a guaranteed buyer and facilitating their borrowing capacity to fund their purchases of Nvidia's AI chips.

Nvidia would earn revenue from the sale of the hardware and a portion of its customers' cloud revenues generated from Nvidia-powered capacity.

However, investor concerns have grown recently as Nvidia continues to inject capital into the AI ecosystem, raising doubts about its involvement in "circular deals" that may artificially boost demand. Earlier this month, Nvidia facilitated $500 billion in financing from major U.S. financial institutions for its customers and guaranteed up to $105 billion to support OpenAI in leasing a substantial data center.

Certain Nvidia employees raised concerns to clients and prospective customers about the potential antitrust scrutiny surrounding the initiative. There are also sensitivities regarding Nvidia's control over how its customers operate. In the program's initial stages, Nvidia's level of control raised objections from some potential partners.

The report stated that Nvidia insisted on supplying its chips to approved customers only and preferred the capacity distribution among multiple smaller firms rather than a single large customer. Under the proposed deals, Nvidia would receive 50% of any revenue earned by cloud providers from its chips above a specified threshold.