虎嗅

When Wall Street Becomes NVIDIA's Sales Channel

原文:当华尔街成为英伟达的销售渠道

Summary of Key Points

NVIDIA, in collaboration with six top Wall Street financial institutions including Apollo, BlackRock, and Berkshire Hathaway, has established an AI computing infrastructure financing platform. The plan is to leverage over $500 billion in third-party capital over the long term to assist customers in building AI facilities (purchasing GPUs, constructing data centers, and establishing power networks). This initiative is not funded by NVIDIA itself; instead, it aims to transform Wall Street into a “sales support channel” to help address the issue of high costs for customers, thereby making financing capabilities a new competitive advantage for chip companies. However, this raises questions: How much of the demand in these orders is genuine? Could Wall Street become a filter that rejects unviable projects?

1. Where Does the $500 Billion Come From? Don’t Let the Numbers Mislead

Many people assume that NVIDIA’s financial strength is reflected in this amount, but that’s not the case:

  • It’s Not NVIDIA’s Money: NVIDIA has neither raised this capital nor committed to contributing any of it.
  • Third-Party Capital: The six financial institutions are operating independently, and the $500 billion represents the total external funds to be mobilized over time. It’s still unclear which projects will be funded first, whether equity or loans will be used, and who will bear the losses if there are any.
  • The Role of NVIDIA: NVIDIA acts as a intermediary, connecting financial institutions with capital with customers in need of AI computing resources, enabling them to access long-term financing at lower costs.

2. Why Does NVIDIA Help Customers Find Funds? A New Approach to Chip Sales

In the past, chip companies focused solely on product performance, software ecosystems, and delivery times. Now, another factor has emerged: the high cost of building AI facilities:

  • The Cost of AI Facilities: Buying GPUs is just part of it; additional investments are required for data centers, power infrastructure, and networking, which take years to pay off, while GPUs become obsolete in just two to three years due to rapid technological advancements. If customers funded these expenses on their own, the recovery period would not align with the speed of technology updates, preventing them from making large purchases.
  • Analogous to Automotive Finance: Chip companies no longer just sell chips; they also offer loans, allowing customers to purchase with a down payment. By helping customers secure funding, NVIDIA enables them to build AI facilities and repay the debt over time, thus facilitating more sales.
  • Financing as a New Selling Tool: In the future, chip companies will compete not only on performance, ecosystems, and delivery but also on their ability to provide affordable and long-term financing.

3. Are the Orders Still Credible with Financing Involved? The Authenticity of Demand

Previously, customers’ willingness to pay for GPUs indicated a genuine interest in the product. Now, with NVIDIA’s assistance in securing funding, orders include elements of supplier credibility and financial leverage. However, this requires closer scrutiny:

  • Don’t Rely Solely on Contract Amounts: For example, if a customer signs a $1 billion GPU order but funds it through Wall Street loans, the success of the project depends on their ability to generate revenue from using the computing power (e.g., by providing services to model companies) and whether they can cover loan costs and equipment depreciation.
  • A Practical Example: If a customer buys 1,000 GPUs with borrowed funds but uses only 30% of the computational capacity, resulting in insufficient profits to cover interest, the order may be unviable, leading to wasted resources.

4. Wall Street Is Not Just a Passive Participant; It Plays a Critical Role

Don’t assume that Wall Street will provide funding without due diligence:

  • Wall Street Filters Projects: The six financial institutions will evaluate the feasibility of each project, considering factors such as the solidity of customer contracts and the efficiency of resource utilization. They won’t invest just because of NVIDIA’s statements.
  • Professional Investors Drive Practicality: Their involvement ensures that AI infrastructure projects are more well-designed, avoiding unnecessary purchases driven by mere chip demand.
  • NVIDIA Does Not Guarantee Losses: NVIDIA merely acts as a intermediary; it does not assume responsibility for customers’ loan repayments. Therefore, the $500 billion should not be seen as NVIDIA’s guarantee of success.

5. The Impact on Chip Competition

The competition among chip companies will expand beyond mere performance and delivery times:

  • Competition in Comprehensive Capabilities: Success will depend on the ability to integrate GPUs, data centers, power infrastructure, and customer contracts into viable projects that attract financing from financial institutions.
  • Key Metrics for Evaluation: Three key factors will determine the success of these projects: the actual cost of financing (interest rates), the resale value of old equipment after GPU obsolescence, and the profitability of the computing centers.
  • NVIDIA’s Advantage: If these metrics are favorable, Wall Street could become a powerful channel for NVIDIA. Otherwise, the so-called “financiable computing power” might merely serve to mask the pressure of selling chips.

In summary, NVIDIA’s strategy transforms chip sales from a product-based model to a service-oriented one that includes financial services, positioning it as a coordinator of the AI ecosystem. However, the success of this approach hinges on whether Wall Street’s funds are directed towards projects with genuine demand. After all, no amount of money can fill an unfeasible project.

---

This translation maintains the structure and tone of the original Chinese analysis while adapting the language to fit financial journalism standards. It also ensures that technical terms are used accurately and consistently throughout the text.