Breaking

NVIDIA to Build AI Factories With Cloud Partners Under Revenue-Sharing Model

July 2, 2026 at 02:55 EDT

Infra & Chips
Sales & Business

On July 1, 2026, NVIDIA announced it will deploy large-scale, multi-tenant "AI factories" in partnership with AI cloud providers. It is introducing a new economic model combining revenue-sharing and credit-support to broaden access to large-scale compute for startups and AI-native companies.

July 1, 2026 · NVIDIA

NVIDIA Becomes a Backer of Its Own Demand

A new "AI factory" model pairs hardware sales with revenue sharing and credit support — helping cloud operators finance capital-heavy compute and giving AI-native firms immediate access at scale.

3-part

Model: hardware sales + ongoing cloud revenue share + credit support

First named partners: Sharon AI & Firmus

6-yr

Sharon AI's strategic compute collaboration with NVIDIA

GPU Scale of the First Two Deployments

Peak GPU count — proportional columns

40,000

Sharon AI

Australia · GB300

170,000

Firmus

Batam Is., Indonesia

Sharon AI · Power

+72MW → 132MW total

Firmus · Power

360MW-scale campus

How the Money Flows

NVIDIA
Sells hardware + backstops capacity, supports credit

→

Cloud Operators
Procure & run AI factories, sell cloud services

→

AI-native firms
Startups, enterprises & ISVs get compute

↺ NVIDIA earns hardware revenue plus an ongoing share of the cloud revenue the capacity generates.

The Bull Case

Lowers the capital barrier for emerging AI firms, unlocks fast large-scale compute for high-volume inference, and is framed as a turning point for sovereign, competitive AI.

The Risk

NVIDIA edges toward a bank-like role underwriting its own demand — raising balance-sheet exposure from unused capacity and doubts over reaching full scale.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.