The pause signals growing internal caution at Nvidia about how far it can go in using its balance sheet and market position to engineer demand for its own chips. Coming so soon after the company also scaled back a proposed financial backstop for OpenAI’s Ohio data centre project, this reinforces a pattern of Nvidia stepping back from arrangements that could expose it to antitrust scrutiny or investor concern over concentrated liability. With $36 billion already committed under the programme’s six year agreements, any prolonged pause or restructuring could affect how markets price the durability of Nvidia’s newer, non-hardware revenue streams, even as the company frames this as normal evolution of a still-developing initiative.
—
Yesterday:
- What Nvidia Earnings Teach Investors About P/E, PEG, Margins and Growth
- Nvidia’s stock does the limbo then the moonwalk on 2028 guidance
—
Summary:
- Nvidia has paused some deals under its AI Compute Partnership, a July financing initiative offering credit support to AI cloud providers in exchange for a share of revenue, per the Wall Street Journal.
- Employees reportedly flagged antitrust concerns and unease over how much control Nvidia exerted on customers’ business practices.
- Partners objected to restrictions on which customers they could rent chips to, and to Nvidia’s preference for spreading capacity across smaller firms.
- Nvidia disclosed $36 billion in commitments under the programme in its latest quarterly filing, with agreements typically running six years.
- Nvidia says its broader July business model remains in place and continues to evolve amid high demand.
Nvidia has paused some deals under a financing programme that offered credit support to artificial intelligence cloud providers in exchange for a share of their revenue, the Wall Street Journal reported, citing people familiar with the matter. The move comes less than two months after the chip giant announced the initiative, known as the AI Compute Partnership.
Some Nvidia employees reportedly raised concerns, both internally and with customers, that the programme could invite antitrust scrutiny, alongside broader sensitivities about how much control Nvidia should exercise over customers’ businesses. Nvidia had told some providers they could only rent chips to approved customers, and indicated a preference for spreading capacity across several smaller AI firms rather than one large customer, a restriction some providers resisted. The precise reason for the pause could not be established, and Nvidia may eventually revamp or restructure the programme.
The initiative was designed to help smaller cloud providers, who must spend heavily on Nvidia chips and data centres before securing firm customer contracts, by having Nvidia pledge to rent unused capacity itself. Sharon AI and Firmus Technologies were named as the first participants when the programme launched in July.
Nvidia disclosed the scale of the programme for the first time this week, revealing $36 billion in commitments under agreements typically running six years, a figure expected to shrink as providers find other customers. CFO Colette Kress told investors the arrangement could generate billions of dollars for Nvidia over the medium to long term. Under the proposed structure, Nvidia and each provider would set a base hourly rate covering costs, with Nvidia collecting 50% of revenue above that threshold.
The pause follows Nvidia’s recent decision to scale back a proposed financial backstop for OpenAI’s Ohio data centre project, suggesting a broader recalibration of how closely Nvidia ties its balance sheet to customers’ fortunes. A company spokeswoman said the underlying business model introduced in July remains in place and continues to evolve amid high demand.
This article was written by Eamonn Sheridan at investinglive.com.