NVIDIA (NVDA) pauses revenue-sharing deals with AI cloud companies and is rethinking the program less than two months after announcing it, according to WSJ

A rapid reversal of a newly announced commercial program is a rarer signal than the program itself: companies at the centre of a capex cycle have historically only unwound go-to-market structures this quickly when internal accounting, legal, or counterparty concerns surfaced after launch, or when the economics proved mispriced on contact with the market. Revenue-sharing arrangements between a dominant chip supplier and the cloud firms that are also its largest customers sit squarely in the debate over circular financing in the AI buildout, where the line between demand creation and demand support has been the central bear case. In past episodes of this kind, the first question from holders is whether the pause is a redesign of terms or a quiet retreat, and the second is whether it touches reported revenue quality, since deals booked with attached equity or revenue give-backs alter how the market capitalises the growth. Worth watching is any disclosure around how much revenue was tied to such structures, commentary from the cloud counterparties involved, and whether the next earnings call reframes the program as shelved rather than paused. Peer and customer stocks with similar entangled arrangements have tended to trade in sympathy on headlines of this type.