Quick Summary
- Nvidia has rolled out a new Nvidia revenue sharing program designed to help fast growing AI startups get their hands on scarce computing power without needing a mountain of cash upfront.
- Startups receive token credits under the program instead of paying full price upfront for GPU access.
- Two initial partners, Sharon AI and Firmus Technologies, are offering potential access to more than 200,000 Nvidia GPUs combined.
- Firmus is building a data center in Batam, Indonesia, expected to scale to 360 megawatts and house up to 170,000 GPUs.
- The move fits a wider pattern of chipmakers and AI firms trading equity, revenue, and compute access to get around funding gaps.
- Nvidia disclosed the plan the same month it moved to raise at least 20 billion dollars through its first bond sale in five years.
- Analysts see programs like this as a way for Nvidia to widen its customer base while tying young AI companies more closely to its hardware.
The chipmaker announced the program this week, framing it as a way for cloud based AI firms and other enterprises to trade future profits for direct access to full stack computing built on Nvidia hardware.
For founders running lean teams, that trade could be the difference between shipping a product and sitting on a waitlist. GPUs are one of the tightest resources in tech, and Nvidia’s new structure changes who gets to use them and how they pay for it.
This piece walks through what Nvidia actually announced, who the first partners are, and what the deal means for startups trying to build on top of Nvidia hardware. It also looks at where this fits into the wider trend of chipmakers and AI companies trading equity and revenue for access to scarce compute.
The Deal Nvidia Is Offering
Under the new partnership program, Nvidia is issuing token credits to fast growing AI startups so they can power their development work. In return, those companies agree to share both product revenue and cloud revenue with Nvidia as their businesses grow.
Nvidia is positioning itself less like a hardware vendor and more like a broker, connecting startups directly to full stack computing capacity that would otherwise be locked up by larger, better funded customers. Instead of one company paying cash for chips, everyone in the chain, from the data center operator to the startup running its models, ends up tied to Nvidia’s ongoing revenue.
Why GPU Access Has Become the Real Bottleneck
Nvidia’s dominance in AI chips has created a strange side effect. Some of the most promising AI companies in the world cannot get enough compute, even when they have funding, because supply simply has not kept pace with demand. GPUs have been compared to oil in how they are traded, and some contracts around them now resemble futures agreements as buyers try to lock in access and manage cost swings.
That scarcity has pushed AI companies toward creative financing. Rather than buying chips outright, many are now trading equity, revenue, or long term purchase commitments just to secure a place in line. Reports on AI industry’s web of circular deals describes this pattern as a way for suppliers, builders, and customers to lock arrangements together so demand for computing power gets met, even if it means blurring the line between investor and customer.
Meet the First Two Partners
Nvidia named two companies that will supply the compute behind the new program. Sharon AI, based in Australia, plans to deploy up to 40,000 Nvidia GPUs as part of the arrangement. Singapore based Firmus Technologies is going even bigger, building a data center in Batam, Indonesia, that is expected to scale to 360 megawatts and eventually house up to 170,000 Nvidia GPUs.
Combined, the two partners are offering potential access to more than 200,000 GPUs, a figure that gives a sense of how much capacity Nvidia is trying to unlock through this model. Sharon AI has previously described its mission as delivering sovereign, large scale AI infrastructure to enterprise clients, startups, and academic researchers who might not otherwise reach that scale of computing power on their own, according to reports.
How the Nvidia Revenue Sharing Deal Works
The mechanics are straightforward once you break them down. Cloud partners like Sharon AI and Firmus build large scale data centers using Nvidia hardware. Startups and other enterprises then tap into that infrastructure using token credits rather than paying the full cost upfront.
As those startups generate revenue from products built on top of that compute, a share flows back through the cloud partner to Nvidia. This gives Nvidia two income streams instead of one. It still sells the underlying hardware, and it now also collects an ongoing cut tied to how much value startups create using that hardware over time.
Part of a Bigger Pattern in AI Financing
Nvidia’s move fits a trend that has been building across the AI industry for more than a year. Chipmakers and cloud providers have increasingly taken equity stakes in the AI companies buying their products, tying supplier and customer together in ways that used to be far less common.
OpenAI shows how far this pattern has gone. The company deployed AMD’s Instinct GPUs under a deal that could see it take a stake of up to 10 percent in the chipmaker, according to OpenAI’s own announcement of the partnership. Around the same time, OpenAI signed a 38 billion dollar deal to run workloads on Amazon’s cloud using hundreds of thousands of Nvidia chips.
Analysts call these arrangements circular deals, since the same handful of companies keep showing up on both sides of the transaction, as suppliers and as investors. Nvidia’s revenue sharing program extends that logic down to a smaller scale, reaching startups too early stage for billion dollar funding rounds but still hungry for compute.
The Trade-Off for Founders
If you run an AI startup, the practical upside here is real. Compute cost and availability have been two of the biggest obstacles between a good idea and a shipped product, and this model chips away at both. You are not writing a massive check before you have proven anything works.
That said, nothing here is free. A portion of your future revenue will flow back through your cloud provider to Nvidia, so the arrangement only makes sense if your business model can support that cost. Read the terms closely before signing, and know exactly how much upside you are trading for early hardware access.
The Debt Raise Sitting in the Background
It’s worth noting the timing here. Earlier this month, Nvidia moved to raise at least 20 billion dollars through its first corporate bond sale in five years, a deal that reportedly drew far more investor demand than the company initially sought. Bloomberg’s coverage of Nvidia’s bond offering noted that Nvidia planned to use part of the proceeds for general corporate purposes, including repaying and refinancing existing debt.
That debt raise and the new revenue sharing program are not the same thing, but they point in a similar direction. Nvidia is leaning on multiple financial tools at once to keep funding the buildout of AI infrastructure while extending its reach into smaller, less established corners of the market.
Final Thoughts
Nvidia’s revenue sharing model marks a real shift in how the company grows its business. Instead of relying only on upfront hardware sales, it is now building financial arrangements that tie its own success directly to the success of the startups and cloud providers using its chips. Sharon AI and Firmus offer the first glimpse of what that looks like in practice, with more than 200,000 GPUs of potential capacity already on the table.
For AI startups that have been priced out of serious compute, this kind of deal could open a door that was previously shut. At the same time, it deepens Nvidia’s grip on the companies shaping the next wave of artificial intelligence. Whether this becomes a standard path into the market or stays a tool reserved for a select group of partners will depend on how the terms hold up once more startups actually sign on.
Discover how AI is reshaping technology, business, and healthcare—without the hype.
Visit InfluenceOfAI.com for easy-to-understand insights, expert analysis, and real-world applications of artificial intelligence. From the latest tools to emerging trends, we help you navigate the AI landscape with clarity and confidence
