Nvidia has halted some transactions under a July-launched model that provides credit support to AI cloud companies in exchange for a share of cloud revenue generated on its chips, the Wall Street Journal reported, after employees raised antitrust concerns about customer leasing and capacity allocation. The step-back occurred last week, less than two months after the program was unveiled, but Nvidia has not scrapped the initiative and may restructure it or fold it into another program. The arrangement was intended to help smaller AI cloud providers finance Nvidia GPU purchases, with Nvidia potentially acting as a guaranteed buyer of unsold capacity and taking half of cloud revenue above a base threshold; the company said on its earnings call that the model could contribute tens of billions of dollars over the medium to long term. Reports that Nvidia had shelved some revenue-sharing agreements pressured neocloud stocks, with Nebius shares falling 4.3% in morning trading after a more than 7% rally on August 27, while CoreWeave and IREN also traded lower; Nvidia shares held their recent range as the report emerged. Nebius’s $5.75 billion convertible-notes offering had closed above its original $4.5 billion target, but investors remained cautious about its planned $20 billion to $25 billion in capital spending. The partial pause comes amid scrutiny of Nvidia’s broader financing role, including facilitating $500 billion in customer credit from major U.S. financial institutions and backing up to $105 billion to help OpenAI lease a large data facility, alongside concerns about circular financing across the AI ecosystem.