Nvidia Has Put Some AI Cloud Deals on Hold Under a New Funding Program — WSJ

Nvidia, a manufacturer of artificial intelligence processors, has put some transactions under its new financing program on hold; the program offered credit support to cloud service providers in exchange for a share of their revenue, The Wall Street Journal reports, citing sources. This allowed Nvidia’s partners to receive funds and use them to purchase the company’s chips.

Nvidia backed away from the program last week, less than two months after its announcement, sources told the WSJ. The exact reason for this decision is unknown, the newspaper noted. In the future, Nvidia may refine the idea or combine it with another initiative, according to some of the publication’s sources.

Some Nvidia employees warned current and potential customers that the program could attract the attention of antitrust authorities and raised the sensitive issue of the extent to which the chipmaker could dictate their terms of business, sources told the newspaper.

The suspension came amid growing criticism of Nvidia for using its financial resources to support projects that subsequently generate demand for its chips, according to the WSJ. The company also recently scaled back its financial support for OpenAI’s large-scale data center project in Ohio due to concerns about how investors would react to the potential liabilities, the publication notes.

Nvidia shares were down 0.6% in the first few minutes after the market opened on August 28. Year-to-date, they are up 22%. Trading the previous day ended with an 8.7% gain, as investors reacted positively to the company’s earnings report and its forecasts.

Nvidia launched a program called the AI Compute Partnership in July to address the financial challenges faced by small cloud providers, according to the WSJ. Building AI clouds requires such companies to invest billions in Nvidia graphics processing units (GPUs) and data centers, often before securing the firm customer contracts needed to obtain loans, the publication explains. Nvidia sought to bridge this gap by promising to lease the GPU capacity itself if the provider could not find another customer. This commitment provided the provider with a guaranteed source of revenue, making it easier to raise the funds needed to build the infrastructure, the WSJ notes.

In its quarterly report this week, Nvidia disclosed the scope of the program for the first time. The company reported that it has committed $36 billion under agreements that typically run for six years, but noted that as providers sell capacity to other customers, this amount will decrease. According to the company’s statement, Sharon AI and Firmus Technologies were the first two cloud providers to participate in the financing initiative.

In the program’s first few weeks, Nvidia drew criticism from some potential partners over the extent of the control it sought, sources told the WSJ. According to them, Nvidia informed a number of providers that they could lease chips only to approved customers. The company made it clear that it preferred to distribute computing power among several smaller players in the AI sector rather than lease it to a single large customer. Some cloud providers objected, arguing that they should be free to choose their own customers, the publication reports.

This article was AI-translated and verified by a human editor