Nvidia Paused Its $36 Billion Cloud Financing Programme After Its Own Staff Raised Antitrust Concerns

Nvidia has paused parts of a financing programme it announced only in July, according to a Wall Street Journal report. The programme, called the AI Compute Partnership, offered credit support to smaller cloud providers in exchange for a share of their revenue. Nvidia disclosed the size of it for the first time in this week’s quarterly filing: 36 billion dollars in commitments, under agreements typically running six years.

The problem, reportedly, came from inside. Nvidia employees raised concerns both internally and with customers that the arrangement could attract antitrust scrutiny, alongside broader unease about how much say Nvidia was taking over its customers’ businesses. Two specific terms drew objections. Nvidia had told some providers they could only rent the chips on to approved customers, and it expressed a preference for spreading capacity across several smaller AI firms rather than concentrating it with one large one. The commercial structure was unusual too: Nvidia and each provider would agree a base hourly rate covering costs, and Nvidia would collect 50% of any revenue above that line. CFO Colette Kress had told investors the arrangement could generate billions over the medium to long term. Sharon AI and Firmus Technologies were named as the first participants at launch. An Nvidia spokeswoman said the underlying July business model remains in place and continues to evolve amid high demand, and the Journal reported the programme may be restructured or folded into something else later.

What’s actually going on here: the idea behind the programme was to solve a real chicken and egg problem. A small cloud provider has to buy racks of very expensive chips and build a data centre before it has signed customers to fill them, which is a terrifying way to spend money. Nvidia offered to remove that risk by promising to rent the unused capacity itself. The catch is what that does to the market. When the company selling the chips also underwrites the buyers, guarantees their revenue, and gets a say in who they may sell to, it starts to look less like a supplier and more like a landlord of the whole neighbourhood. That is precisely the shape regulators pay attention to. It also fits a pattern: Nvidia recently scaled back a proposed financial backstop for OpenAI’s Ohio data centre project too.

What this means for you: nothing directly, and that is worth saying plainly. This will not change the price of your chatbot subscription this month. But if you follow AI at all, this is a useful window into how much of the current boom is financed rather than earned. A meaningful chunk of the compute capacity being built right now exists because the chip vendor promised to be the customer of last resort. When even Nvidia’s own employees start asking whether that is too much control in one company’s hands, it tells you something about how tightly wound the supply chain has become, and how much of the growth story depends on arrangements that could be unwound.

Sources

• WSJ report: Nvidia pauses AI cloud revenue share deals amid antitrust and control concerns (investingLive)
• Nvidia pauses revenue sharing deals with AI cloud companies, WSJ reports (Reuters via Yahoo Finance)
• Nvidia pauses AI cloud revenue sharing deals over antitrust concerns (Yahoo Finance)