Nvidia is increasingly financing the companies that buy its chips, creating a capital cycle that could help fuel the next phase of the AI infrastructure boom.
The company has committed nearly $50 billion to AI laboratories and has lined up partnerships intended to facilitate more than $500 billion in outside financing for AI infrastructure. Nvidia’s CFO Colette Kress said the companies backed by Nvidia could account for roughly a quarter of its business next year.
The structure has attracted attention because of its resemblance to circular financing: Nvidia provides capital or financial support to AI companies, those companies use the resources to build data centers, and those facilities purchase Nvidia’s GPUs.
How the financing cycle works
The basic mechanism is relatively straightforward.
Nvidia invests in or provides financial support to an AI company. That company then uses the capital to expand its computing infrastructure, purchasing large quantities of Nvidia hardware.
Those purchases generate revenue for Nvidia, strengthening its balance sheet and potentially increasing the value of its investments. Nvidia can then deploy additional capital into the AI ecosystem.
The result is a feedback loop connecting Nvidia’s financing activities with its chip sales.
Nvidia is working with major financial institutions
Nvidia has partnered with six major investment firms to create financing platforms designed to raise more than $500 billion in outside capital for AI infrastructure.
The firms named by Kress include Apollo Global Management, BlackRock, Blackstone, Brookfield Asset Management, Goldman Sachs and KKR.
The arrangements are intended to help AI companies overcome one of the biggest barriers facing the industry: the enormous upfront cost of building data centers.
Nvidia has also worked with SB Energy to secure land, electricity and construction capacity for facilities designed around Nvidia hardware. The first phase is expected to support 4.25 gigawatts and will be used by OpenAI.
OpenAI’s existing and planned commitments represent approximately 12 gigawatts of Nvidia computing capacity through 2030.
Why Nvidia says this is not circular financing
Nvidia itself has acknowledged that some investors may describe these arrangements as circular financing, but the company disputes that characterization.
Kress emphasized that outside lenders continue to evaluate individual transactions and that Nvidia is not directly making the loans.
The company also points to the credit quality of the customers purchasing its equipment. Nvidia says its chips are being sold to investment-grade customers or companies backed by investment-grade entities.
Another factor is the resale value of Nvidia hardware. If a customer were unable to meet its obligations, Nvidia argues that the equipment could potentially be transferred to another customer.
That assumption is important because Nvidia’s exposure would be significantly greater if the hardware became difficult to sell.
AI companies face a massive infrastructure problem
The financing arrangements highlight a fundamental challenge facing the AI industry.
AI companies may have enormous demand for computing but lack the long operating history, long-term contracts and credit ratings that traditional lenders typically want to see before financing large data center projects.
That creates a financing gap.
Nvidia and financial institutions can help bridge that gap by providing investors with additional confidence that the infrastructure will generate revenue.
For Nvidia, the strategy also helps ensure that AI companies have access to the hardware necessary to continue expanding.
Nvidia is betting heavily on AI agents
The company’s infrastructure expectations are also tied to the increasing use of AI agents.
Kress said an AI agent can require between 15 and 100 times the computing power of a person performing the same task. Nvidia CEO Jensen Huang has similarly argued that AI has recently shifted toward predominantly agentic workloads.
If AI agents become significantly more widespread, demand for computing could increase dramatically.
That would benefit Nvidia because more AI workloads require more GPUs, networking equipment and data center infrastructure.
The company expects demand to remain strong
Nvidia expects to generate approximately $108 billion in revenue during the current quarter and has indicated that it expects annual growth of around 70% through January 2028.
Kress said the primary constraint is increasingly supply rather than demand.
However, the rapid expansion of AI infrastructure is also creating pressure elsewhere in the supply chain.
Memory prices, for example, are rising faster than Nvidia previously anticipated. The company expects gross margins to decline to approximately 74% before potentially reaching a low of around 71% to 72%.
The biggest question is what happens if demand changes
The financing strategy creates an important risk for Nvidia and its investors.
If AI demand continues growing faster than available computing capacity, the model could work extremely well. AI companies receive the financing they need, data centers are built, Nvidia sells more equipment and financial institutions earn returns on their investments.
But the economics could look very different if AI infrastructure becomes overbuilt.
If an AI company struggles financially, Nvidia could potentially face losses on its investment while simultaneously losing a major customer for its chips.
The company’s argument that its equipment can simply be sold to another customer depends heavily on continued demand for AI computing.
A powerful but closely connected ecosystem
Nvidia’s financing strategy illustrates just how interconnected the AI infrastructure market has become.
The company is no longer simply selling GPUs to technology companies. It is increasingly involved in the broader ecosystem responsible for financing, constructing and operating the infrastructure that uses those GPUs.
That creates an enormous opportunity if AI adoption continues accelerating.
At the same time, it means Nvidia’s financial exposure to the AI ecosystem is becoming more complicated. Its investments, partnerships, financing arrangements and hardware sales are increasingly tied to the same underlying demand for AI computing.
The strategy could ultimately help Nvidia maintain its dominant position as AI infrastructure expands, but it also means investors will be watching closely to determine whether the enormous spending cycle is supported by sustainable demand.
Source: https://www.artificialintelligence-news.com/news/nvidia-circular-financing-ai-labs/
