Nvidia Suspends AI Cloud Revenue-Sharing Plan Amid Antitrust Concerns
nashnova research
Nvidia last week suspended parts of its AI Compute Partnership revenue-sharing agreements, less than two months after launch, after internal staff flagged potential antitrust risk; the pause signals that Nvidia's attempt to earn twice — first on chip sales, then on cloud rental income — hit a regulatory tripwire before it could scale.
What was the plan trying to do?
Nvidia launched the AI Compute Partnership in July with a double-dip revenue model: sell GPUs to cloud providers, then take a cut of the rental income those providers earn from leasing the chips to end customers.
The split works like this: Nvidia and the cloud provider set a base hourly rate covering the provider's costs — chip depreciation, data-center operations, staffing. Revenue above that baseline is split 50/50.
In plain terms = Nvidia no longer just sells the shovels — it wants half the gold that gets dug up with them.
Why did it suddenly stop?
According to the Wall Street Journal, Nvidia's own employees raised concerns first, warning the model could invite antitrust scrutiny.
The sensitive point: Nvidia is both the dominant chip supplier and the party dictating how customers operate and how revenue is split. The monopoly optics are hard to avoid.
Nvidia's spokesperson said the plan "remains operational" and is "continuing to evolve," but did not confirm the exact reason for the pause.
Why did partners push back?
Nvidia required cloud providers to lease chips only to Nvidia-approved customers and preferred spreading capacity across many small AI firms rather than renting to a single large client.
Some providers objected outright, arguing that choosing their own customers is a basic right.
This reflects a fundamental tension: Nvidia wants control to protect its revenue share, but partners refuse to act as contract assemblers.
How big is this program?
Nvidia disclosed the scale for the first time this quarter: $36 billion in commitments, with agreements typically spanning six years.
CFO Colette Kress told investors this week that the new revenue stream could generate billions of dollars over the medium to long term.
This means → even a "partial pause" involves sums large enough to reshape how the market assesses Nvidia's business-model pivot.
What else is Nvidia pulling back on?
The pause is not an isolated move — Nvidia also scaled back its financial-backstop plan for OpenAI's large Ohio data-center project, partly over concerns about investor reaction to potential liabilities.
People familiar with the matter say the revenue-sharing plan may return in a revised form or be folded into other programs.
The core question remains: can Nvidia rebuild this "sell chips + share rent" double-profit model without tripping regulatory red lines? The market is still watching.
市场有风险，内容仅供研究参考，不构成投资建议。