Nvidia is making a colossal bet on the future of artificial intelligence, funneling nearly $50 billion into AI labs that are its primary customers and securing commitments for an astounding $500 billion in future investment.
This aggressive strategy, detailed by Chief Financial Officer Colette Kress, signals a deeply integrated approach to fueling AI development. Kress indicated that demand from the very AI labs Nvidia actively supports will account for approximately a quarter of the company’s business next fiscal year. This arrangement, which Nvidia itself has termed “circular financing,” is a nuanced play designed to accelerate the AI ecosystem while bolstering Nvidia’s own market position.
The mechanics of this “circular financing” are straightforward yet powerful. Nvidia injects capital or facilitates credit into emerging AI labs. These labs then leverage this financial backing to construct data centers, which are invariably outfitted with Nvidia’s cutting-edge chips. The subsequent purchase of these chips is recognized as revenue for Nvidia, thereby expanding its financial resources and market capitalization, which in turn fuels further investment into the AI sector. It’s a self-reinforcing cycle designed to drive exponential growth across the AI landscape.
Nvidia’s Ambitious Commitments Unveiled
The scale of Nvidia’s commitment is staggering. Kress revealed that the company has forged partnerships with six prominent investment firms – Apollo Global Management, BlackRock, Blackstone, Brookfield Asset Management, Goldman Sachs, and KKR. These collaborations are aimed at establishing financing platforms that will mobilize over $500 billion in external capital, specifically earmarked for AI labs to fund their ambitious data center build-outs.
Beyond financial partnerships, Nvidia is also actively securing critical infrastructure. Kress highlighted an agreement with SB Energy to secure land, power, and building capacity exclusively for Nvidia equipment. The initial phase of this infrastructure is designed to support 4.25 gigawatts of computing power and will be utilized by OpenAI, with Kress noting OpenAI’s existing and planned commitments to utilize approximately 12 gigawatts of Nvidia compute power through 2030. In a separate arrangement, Nvidia is providing credit support for nearly two gigawatts of capacity for an undisclosed second AI lab.
It’s important to note that these “partnerships” are described as being “subject to definitive agreements,” meaning the binding contracts are still in development. The $500 billion figure represents a significant intention and a clear signal of future market direction, rather than immediately deployed capital.
Nvidia is also extending its influence to smaller cloud operators by guaranteeing a portion of their capacity. This provides lenders with a predictable revenue stream, making it easier for these operators to secure financing for their infrastructure. In return, Nvidia participates in the operator’s revenue above a certain threshold, effectively securing multiple revenue streams from its own hardware and the subsequent rental income generated by the cloud operator.
Navigating the “Circular Financing” Label
Nvidia is keen to distinguish its strategy from traditional “circular financing.” Kress articulated three key reasons for this distinction:
Firstly, she emphasized that external lenders independently assess the creditworthiness of each deal, and Nvidia is not directly issuing loans. The chips are supplied to customers who are either investment-grade or backed by such entities. Furthermore, she noted that in the event of a customer default, the hardware can be redeployed to another buyer, mitigating Nvidia’s direct financial exposure.
Secondly, Kress explained the critical need for this support. Many AI labs face a demand for computing power that outstrips their current financial capacity. As relatively young companies, they often lack the long-term contracts and established credit ratings that traditional lenders require for data center financing. Their growth is primarily constrained by access to compute, not by market demand or technological innovation.
The inherent risk in such a model is the potential for a customer to default. This would result in both a lost sale and a write-off of the initial investment for Nvidia. Kress’s counterargument is that the hardware is highly mobile and can be resold, a scenario that holds true as long as demand for Nvidia’s chips continues to outpace supply, which the company asserts is currently the case.
When questioned by analysts about Nvidia funding labs that are developing their own silicon, such as OpenAI’s exploration into custom processors, CEO Jensen Huang responded that Nvidia offers a versatile platform applicable across various cloud environments and the entire lifecycle of an AI system. He contrasted this with proprietary chips designed for single-use cases. Reflecting on his investment strategy, Huang expressed a singular regret: not investing more, sooner.
The Underlying “Agent” Assumption
The strategic underpinning of Nvidia’s vision appears to be the escalating role of AI “agents.” Kress noted that an AI agent can require anywhere from 15 to 100 times the computing power of a human user. Huang further elaborated that AI has recently transitioned to a predominantly “agentic” paradigm, though Nvidia has not provided specific data to quantify this shift.
Based on this agent-driven growth, Nvidia has projected $108 billion in revenue for the current quarter. Looking ahead, the company anticipates approximately 70% growth for the fiscal year ending January 2028, a figure Kress indicated is primarily constrained by supply chain limitations rather than market demand.
In a separate development, Kress warned that memory prices are increasing at a faster pace than anticipated. Consequently, Nvidia has revised its gross margin guidance downward to 74% for the current quarter, with expectations of it bottoming out between 71% and 72% in the fourth quarter. The scarcity of memory, she explained, is largely a consequence of the accelerated AI infrastructure build-out.
Nvidia is scheduled to release its next earnings report on November 17. The identity of the AI lab receiving the credit support for nearly two gigawatts of capacity remains undisclosed.
Original article, Author: Samuel Thompson. If you wish to reprint this article, please indicate the source:https://aicnbc.com/25201.html
