Nvidia offers AI startups compute power now in exchange for a cut of future revenue

Nvidia Launches Revenue Sharing Program to Address Financing Gap

Nvidia has introduced a new business model that fundamentally changes how AI startups can access compute power. The Nvidia revenue sharing program allows AI cloud providers to procure GPU infrastructure and sell Nvidia-powered cloud services while the chip giant collects both standard product revenue from hardware sales and a percentage of the AI cloud revenue generated on that capacity 1

1

. In a blog post co-authored by CFO Colette Kress, the company pitched the arrangement as a way to open compute access to AI startups that cannot finance it themselves 1

1

.

The new Nvidia financing model addresses a critical capital problem in the AI industry. Even signed, long-term customer commitments have failed to convince lenders to fund large-scale deployments, leaving smaller cloud providers unable to borrow against the demand they had already generated 1

1

. Under this structure, developers receive token credits in exchange for a slice of their future sales, though neither Nvidia nor its partners has disclosed the specific revenue-split percentages .

Sharon AI and Firmus Technologies Lead Initial Deployment

Australia's Sharon AI and Singapore-based Firmus Technologies are the first named partners in this initiative. In an 8-K filing dated June 12th, Sharon AI disclosed that the agreement runs for six years and covers 72 MW of new Australian data center capacity built to Nvidia's DSX AI factory design, scaling to as many as 40,000 Grace Blackwell GB300 GPUs . James Manning, Sharon AI's cofounder and chief executive, called the deal "a pivotal moment" for the company's push into sovereign, large-scale AI compute 4

4

.

Firmus is building an even larger campus in Batam, Indonesia, that's expected to scale to 360 MW and house up to 170,000 Nvidia GPUs across Grace-Blackwell, Vera-Rubin and Vera platforms 1

1

4

4

. Bloomberg reported that Firmus expects between $25 billion and $30 billion in committed offtake agreements over the deal's first six years 4

4

.

How the Revenue Sharing Model Changes GPU Access

The arrangement gives Nvidia something it has not had at this scale before: a recurring revenue stream layered on top of hardware sales 4

4

. This usage-linked earnings stream ties a slice of Nvidia's income to utilization instead of just hardware sales . If partner clouds cannot keep racks rented, the usage-linked stream shrinks, creating a live concern given the depreciation pressure already building on operators paying off hardware that Nvidia refreshes essentially every year 1

1

.

The model pairs revenue sharing with credit support, effectively helping smaller AI clouds finance the purchase in the first place 4

4

. It functions like vendor financing for AI startups with an equity-like upside attached, though specifics on how this business model will work in practice remain thin 2

2

. Cloud providers such as CoreWeave and Lambda have had to borrow billions of dollars from venture capitalists and hedge funds to bankroll their data center build outs, making this alternative particularly attractive 2

2

.

Strategic Shift from Direct Investment to Revenue Participation

This new model inverts Nvidia's recent investment strategy. The company has spent much of the last year funneling cash directly to its customers, including a $30 billion participation in OpenAI's $110 billion funding round and backing for xAI's $20 billion Colossus 2 financing, arrangements that drew repeated circular financing criticism 1

1

. Rather than investing capital that returns as GPU orders, Nvidia now extends credit support and collects a royalty on its partners' sales for years afterward 1

1

.

Nvidia has already committed more than $40 billion to direct AI equity investments this year, spanning OpenAI, Nebius and dozens of smaller rounds 4

4

. The revenue-sharing compute model does something similar without touching the cap table, keeping the balance sheet exposure with its cloud partners instead of on its own books 4

4

.

Implications for the AI Ecosystem and Compute Capacity

The initiative aims to serve companies that need immediate, elastic access to AI infrastructure for training, fine-tuning and high-volume inference without committing to years of hardware procurement themselves 4

4

. Nvidia named Baseten, Fireworks AI and Together AI as examples of customers this is meant to serve 4

4

. It represents a bet on the long tail of model builders, agent platforms and enterprises that want frontier compute capacity but not the balance-sheet risk of building a data center 4

4

.

The arrangement could provide insulation against a potential AI bust. If demand for new GPUs falls, Nvidia may still earn recurring revenue from the GPUs it has already sold, assuming customer demand remains high 2

2

. However, if AI-native demand cools, Nvidia is now exposed to that slowdown twice: once through chip sales and again through the AI cloud revenue it has agreed to share 4

4

. This deepens a dependency that has already drawn scrutiny, as an increasing share of the AI industry's growth becomes contractually tied to Nvidia's own success 4

4

.