Nvidia announced a new commercial partnership model with AI cloud providers to deploy large-scale, multi-tenant AI factories powered by its DSX platform, enabling massive AI computing in the cloud [1, 2, 3, 4, 5]. The model uses revenue sharing and credit support, helping cloud providers scale capital-efficiently and accelerate AI infrastructure rollout while generating recurring revenues for Nvidia tied to usage beyond just hardware sales [1, 2, 3, 4, 5].
Australian AI firm Sharon AI and Indonesia’s Firmus Energy are the first adopters. Sharon AI plans to deploy 40,000 Grace Blackwell GB300 GPUs, while Firmus aims for about 170,000 GPUs with a 360 megawatt power capacity [4, 5]. These deployments demonstrate the scale of Nvidia’s new AI factory concept.
Nvidia CEO Jensen Huang described the next-generation data centers as “AI factories” where the core shifts from file storage and transmission to producing tokens serving as raw materials for code, answers, designs, and services [3]. Global deployments of Nvidia-based AI factory infrastructure now span nearly 40 countries [3].
Taiwanese hardware manufacturers such as Hon Hai (Foxconn), Quanta, Wistron, and Wiwynn are key suppliers in this ecosystem [3]. Hon Hai is a major partner supporting Nvidia’s expansion and will enhance manufacturing at its Houston plant using Nvidia AI systems. Hon Hai Chairman Liu Yangwei said the world is entering an AI factory era, projecting global data center investments to reach $1.6 trillion by 2030 [3].
Market reactions to AI sector developments have been volatile. Meta’s plan to rent out excess computing resources has weighed on AI stocks recently [2]. Meanwhile, investor Michael Burry, known for predicting the US housing crash, publicly revealed a heavy short position on Nvidia around $198 per share in early July. He warned the current AI enthusiasm resembles a technology bubble, risking rapid hardware obsolescence and aggressive capital expenditure [4, 5]. However, supporters argue strong AI inference demand will drive enterprise adoption and cement Nvidia’s leadership [4, 5].
Nvidia officially announced the AI factory revenue sharing and credit support model in early August 2026 [1, 2, 3]. The rollout continues with new deployments and partnerships expected in the coming months.
