Summary of Key Points
NVIDIA, in collaboration with six top Wall Street financial institutions including BlackRock and Goldman Sachs, has established a financing platform worth over $500 billion. The purpose of this platform is to assist customers within its ecosystem—AI labs, enterprises, and cloud service providers—in acquiring hardware and building “AI factories.” The core idea is to allow customers to avoid using their own cash (thus preventing cash flow shortages), while the financial institutions can earn stable long-term returns by investing in computing power projects. However, some question whether this constitutes a form of “circular financing” (where NVIDIA helps customers purchase its own chips, and the money then flows back to NVIDIA). Jensen Huang responded by stating that independent capital has been brought in, and the demand for computing power is genuine. Whether this model will be successful in the long run depends on whether GPU rental prices continue to rise.
How Does the Financing Platform Work?
The logic behind the platform is straightforward:
- For Clients: Looking to build an AI factory but lacking the funds to purchase NVIDIA GPUs? Use this platform to borrow money. The interest rates are more favorable than those available in the market, and there’s no need to use corporate cash (which does not affect the balance sheet); you can simply repay the loan with the income generated by the future use of computing power.
- For Financial Institutions: By investing in the platform, they receive long-term returns linked to the usage of computing power. For example, the institution receives a portion of the revenue generated by the client’s AI factory each day, effectively turning AI computing power into a stable, long-term, investable asset similar to infrastructure such as electricity or highways.
- NVIDIA’s Role: NVIDIA does not provide the funds directly; it merely acts as the facilitator. It offers a safety net: if a project fails, NVIDIA will cover up to 25% of the losses, with the remaining risks assessed by the financial institutions.
Why the Urgency to Establish This Platform?
The cost of building AI infrastructure is extremely high. Companies like Microsoft and Google have spent $800 billion this year on AI-related investments (including purchasing equipment and building data centers), and that figure is expected to rise to $1.1 trillion next year. If these expenses are funded directly from corporate cash flows, it will significantly reduce their free cash flow (the money earned minus the money spent). For instance, Meta’s free cash flow dropped from $8.5 billion in the second quarter of this year to only $7.8 billion due to increased capital expenditures, which also caused its stock price to fall by 7%. Investors are concerned that these companies might invest all their resources in AI infrastructure and fail to recoup their costs.
NVIDIA’s platform aims to solve this problem for customers—if you don’t have the money to buy our chips, we’ll help you secure a loan from banks, allowing you to continue purchasing while we can continue selling our products.
The Debate Around Circular Financing
Some argue that this is circular financing: NVIDIA invests in customers, who then use the funds to purchase its chips, effectively creating demand for its own products. Short-term fund manager George Soros even suggested that if the AI bubble bursts, NVIDIA and these financial institutions could face similar scrutiny as they did during the 2008 financial crisis.
How does Jensen Huang respond?
- Independent Capital: The funds come from six major financial institutions, which will evaluate the feasibility of each project on their own, not at NVIDIA’s discretion.
- Genuine Demand: There is a real demand for AI computing power. For example, the A100 chip, introduced six years ago, is still in use, and the rental price of the H100 chip has increased this year—indicating a persistent need for computing power.
- Limited Risk: NVIDIA only covers up to 25% of the losses, so it does not expose itself to significant risks.
The Significance of the Platform and Future Challenges
The significance of this platform lies in its transformation of GPUs from mere hardware into investable assets. BlackRock’s CEO compared it to mortgage-backed securities from the 1970s, which packaged individual mortgages into assets for investors to buy, allowing more capital to flow into the real estate market. Now, NVIDIA is doing the same with computing power, enabling more capital to be invested in AI infrastructure.
However, the success of this model hinges on a key indicator: GPU rental prices. When NVIDIA releases its second-quarter financial report on August 26th, if rental prices reach record highs, it will demonstrate that computing power is indeed valuable and the financial institutions’ investments will be profitable, validating the platform’s logic. If prices fall, it will raise concerns that AI infrastructure might be a bubble.
Conclusion
NVIDIA’s financing platform essentially uses financial mechanisms to address the shortage of funds in AI infrastructure, helping customers stay operational while also promoting sales of its chips. Whether it can overcome doubts about circular financing depends on whether there is a sustained real demand for computing power. Ultimately, whether computing power becomes as valuable as essential resources like electricity and water will depend on market willingness to pay for it.