Nvidia Ends Revenue-Sharing Agreements with AI Cloud Providers

Nvidia is discontinuing its revenue-sharing arrangements with AI cloud providers, a decision that could significantly alter the landscape for smaller entities in the artificial intelligence infrastructure sector. This model, which began on July 1, 2026, allowed AI cloud providers to acquire Nvidia’s GPUs with financial backing from the company, in exchange for a share of the revenue generated from these GPUs.

Initial participants in this program included Sharon AI from Australia, which planned to deploy up to 40,000 Grace Blackwell GB300 GPUs over six years, and Indonesia’s Firmus Technologies, with a potential scale of 170,000 GPUs and expected revenues between $25 to $30 billion over the same period.

The revenue-sharing model built on previous agreements with companies like CoreWeave and Lambda, which were valued at $6.3 billion and $1.5 billion, respectively. Analysts referred to this approach as “vendor financing,” a strategy more typical of telecom equipment firms than semiconductor companies.

As of the second quarter of fiscal year 2027, Nvidia’s non-hyperscale segment, known internally as “ACIE,” accounted for about 50% of the company’s data center revenue. The cessation of these revenue-sharing agreements poses a significant challenge for neocloud providers that were either involved in or planning to join these programs, as they will now need to seek alternative funding sources.

Investors are advised to monitor changes in the ACIE segment’s contribution to data center revenue in the upcoming quarters. A notable decline from the current 50% could indicate that the revenue-sharing model was more driven by Nvidia’s financial incentives than by genuine demand from smaller providers.