Particle.news
Download on the App Store

Nvidia Pauses Some GPU Revenue‑Share Deals With AI Cloud Providers

The company says the programme will be reworked after employees and partners raised concerns that its financing terms gave Nvidia too much control and could invite regulatory scrutiny.

Overview

  • Nvidia paused a set of individual deals under its July revenue‑sharing financing programme on Thursday, according to people familiar with the matter and company statements.
  • The company disclosed about $36 billion in multi‑year commitments tied to the programme in its latest quarterly filing, with agreements typically running six years.
  • Current and prospective partners complained that Nvidia had tried to limit who they could rent capacity to and favored spreading capacity across multiple small customers, and some employees warned those terms could draw antitrust attention.
  • Smaller AI cloud providers that had expected Nvidia credit support now face the prospect of finding new capital or reshaping deployment plans, a shift that could change how Nvidia’s non‑hyperscale data‑center revenue is viewed by investors.
  • The programme builds on earlier vendor‑style financing and a recently scaled‑back Ohio backstop for a large data centre, raising questions about concentrated counterparty exposure, circular financing, and hardware obsolescence risks.