Nvidia has paused some of the deals at the center of its novel financing initiative that offered credit support to AI cloud companies in exchange for a share of their revenue, according to a Wall Street Journal exclusive published Thursday. Reuters confirmed the report, noting that the chipmaker is rethinking the program less than two months after announcing it.
The pause marks a striking retreat from one of Nvidia's most aggressive financing experiments, and it lands amid mounting investor scrutiny of the circular deals that have powered the AI infrastructure boom. For readers following the latest AI developments, the move is the clearest sign yet that the vendor-financing model underpinning the "neocloud" gold rush is being recalibrated.
What the Revenue-Sharing Program Was
Announced in early July 2026, the initiative let Nvidia back AI cloud providers — the startups and specialist operators often called neoclouds — by offering credit support and access to Nvidia systems in exchange for a percentage of the cloud revenue those systems generated. Industry reports at the time described deployments of up to 210,000 Grace Blackwell GPUs under the framework, aimed at startups that lacked the capital or balance sheets to buy AI infrastructure outright.
For Nvidia, the arrangement converted one-time chip sales into recurring income tied to cloud usage. For the AI cloud companies, it removed the upfront cost of GPUs — the single largest expense in their business — in exchange for handing Nvidia an ongoing claim on their future revenue.
Why Nvidia Is Rethinking It Now
The Wall Street Journal reported that Nvidia has paused some deals under the program while it reassesses the approach. Neither the company nor the newspaper detailed which specific transactions were affected, but the timing is telling: the pause comes less than two months after the program's launch and follows a string of reports about Nvidia's widening web of financing commitments.
Those commitments have grown enormous. Nvidia has been in talks with Apollo, Blackstone, BlackRock's Global Infrastructure Partners, Brookfield, Goldman Sachs and KKR on a $500 billion funding package for AI infrastructure. It also negotiated — and then sharply scaled back — a guarantee for an OpenAI data center project in Ohio, cutting the planned commitment from roughly $250 billion to under $120 billion after investors balked at the level of risk the chipmaker would absorb.
The common thread is investor unease about circularity: Nvidia finances the customers who buy its chips, inflating demand for its own products while concentrating risk on its own balance sheet. The revenue-sharing program intensified that pattern, since Nvidia would effectively be swapping hardware for equity-like exposure to young cloud operators whose businesses depend entirely on demand for AI computing.
What It Means for the AI Cloud Market
The neocloud sector — companies such as specialist GPU rental providers that sprang up to serve AI training and inference demand — has relied on vendor financing to scale at a pace conventional lending would never support. A pause in revenue-sharing support does not mean Nvidia is exiting those relationships, but it removes a growth channel that was announced with fanfare only weeks ago.
Analysts have warned for months that the financing structures behind the AI buildout deserved closer inspection. A pause by the sector's most important supplier suggests Nvidia itself sees reason for caution, whether from credit concerns about individual partners, from investor relations pressure, or from a desire to direct capital toward larger, more strategic deals.
The Broader Context
Nvidia's financing retrenchment lands alongside other signs that the AI infrastructure race is entering a more disciplined phase. Wall Street banks are being pulled in to share the funding burden. OpenAI's data center guarantees were cut down to size. And Nvidia's own blockbuster revenue projections — the company has forecast up to $673 billion in fiscal 2028 sales — mean the stakes of any misstep are measured in hundreds of billions of dollars.
For now, the company's core business remains extraordinarily strong: data center revenue continues to climb on relentless demand for AI computing. But the pause on revenue-sharing deals is a reminder that even the most powerful company in the AI boom is hedging against the possibility that some of its customers' growth is borrowed rather than earned.
Reuters reported that Nvidia did not immediately respond to a request for comment outside regular business hours.
A Financing Model Under a Microscope
The revenue-sharing framework was part of a broader shift in how Nvidia sells. Rather than simply shipping chips, the company increasingly acts as a lender, lessor and guarantor across the AI supply chain — backing data center projects, investing in its own customers, and in this case trading hardware access for a slice of cloud revenue. Each structure deepens Nvidia's exposure to the fortunes of the companies that depend on its silicon.
That exposure is now large enough that Wall Street analysts treat Nvidia's financing commitments as a systemic question for the AI trade. When the OpenAI Ohio guarantee was slashed from $250 billion to under $120 billion earlier this month, the stated reason was investor concern about risk concentration. Pausing revenue-sharing deals with smaller AI cloud operators extends the same logic one tier down the stack: if a guaranteed commitment to the world's most valuable AI startup invites scrutiny, revenue-share credit support for unproven neoclouds invites more.
There is also a competitive dimension. Rival accelerator vendors and cloud providers have courted the same neocloud customers with their own financing and partnership programs. A pause leaves room for competitors to court startups that were counting on Nvidia-backed deployments, though none can match Nvidia's manufacturing scale or the gravitational pull of its software ecosystem.
Whether Nvidia revives, restructures or quietly retires the program will say a lot about how the next phase of the AI buildout gets financed. If vendor financing gives way to Wall Street capital pools — the $500 billion package reportedly under discussion with Apollo, Blackstone and others — risk shifts from a single chipmaker's balance sheet to diversified investors, which is precisely what the largest infrastructure buildout in history may require to stay solvent.
Reuters reported that Nvidia did not immediately respond to a request for comment outside regular business hours.
Stay Ahead of AI
The money behind AI is moving as fast as the models. For breaking AI news on financing, chips and the companies building the compute era, AI Buzz Wire tracks every turn.
Read more AI news →
