Nvidia has denied a report claiming it placed some transactions under its recently launched AI Compute Partnership on hold, less than two months after unveiling the initiative in early July and just days before it detailed the effort during its earnings call.
According to the report, the deals were paused because some partners were frustrated with Nvidia’s attempts to influence their operations, and because the arrangement raised internal concerns about potential antitrust scrutiny. Nvidia pushed back on that characterization.
“The new business model we introduced in July that opens up compute access to the fast-growing AI ecosystem is still in place and continues to evolve due to high demand,” a company spokesperson said.
The report does not claim Nvidia abandoned the program, under which the company committed to rent capacity from newly built AI data centers and guarantee a minimum revenue level. Instead, it states that certain deals were put on hold. Nvidia’s response confirms the program remains active, though it is evolving.
What the Restrictions Reportedly Involved
The report indicates Nvidia sought to control how its AI Compute Partners rented out their capacity. The company reportedly told some cloud providers in the program that they could lease its GPUs only to customers approved by Nvidia. Nvidia also preferred to distribute available capacity across multiple smaller AI companies rather than let a single large customer claim most or all of it.
Some cloud operators reportedly resisted these conditions, arguing they should keep control over which customers they serve. That pushback may have prompted Nvidia to hold back on some agreements.
Nvidia does not lend money or directly finance AI data center construction, which would amount to circular financing. Instead, it provides demand commitments and guaranteed revenue levels, an arrangement that may have raised the internal antitrust concerns. As a result, the company could be revising the terms of deals it signs with partners.
How the EUR 31 Billion Commitment Works
Modern AI data centers cost billions of dollars that must be spent on facilities, infrastructure, and compute hardware long before an operator secures enough customer contracts to fund construction. Lenders and infrastructure investors want assurance that enough future capacity will actually be rented.
Under the program, Nvidia applies its own demand commitment to a portion of a facility’s capacity in exchange for a percentage of that facility’s revenue when demand is strong. This makes financing easier because part of a project’s revenue stream is effectively backed by Nvidia rather than relying entirely on the operator’s ability to find customers.
The initiative involves EUR 31 billion in commitments, structured on a take-or-pay basis.
Source
Image: tomshardware.com
