Nvidia widens AI customer support as funding loop grows
2026-08-21
Investment and credit support
US chipmaker Nvidia is moving beyond the role of a semiconductor developer and becoming a source of funding for the broader artificial intelligence industry. By providing customers with money to buy chips and credit guarantees, Nvidia is helping create demand itself, a practice that can obscure the true end demand for AI investment and spread risk across the industry if momentum stalls.
Nvidia's funding support falls broadly into two categories. One is direct investment in AI startups and other companies using its abundant cash on hand. The other is support for customers through outside capital. In both cases, the seller itself is helping finance purchases of GPUs needed for AI development, creating a circular flow of money.
Based on an analysis by research firm CB Insights, Nikkei aggregated disclosure data from the past year and found that Nvidia has made at least 24 investments in AI companies and emerging cloud providers, among other customers or potential customers. Even among the seven companies that disclosed investment amounts, the total came to $48.6 billion, or about 7.7 trillion yen.
The biggest deal is a $30 billion investment in OpenAI, the US company behind ChatGPT. Nvidia has also committed up to $10 billion to Anthropic and is funding emerging cloud companies such as CoreWeave. Including startups that have not disclosed their funding, the total is likely to be even larger.
It is difficult to separate these investments from pure financial investments, but many of the recipient companies are expected to direct the money toward securing computing capacity. That means revenue flows back to Nvidia, which holds more than 70% of the global AI chip market. The arrangement resembles vendor financing, a practice seen around the dot-com bubble in the early 2000s, in which suppliers helped customers buy their products.
Outside capital and credit support
In addition to direct investment, Nvidia is also expanding customer support using outside capital. On the 10th, it said it would team up with six major financial firms, including Apollo Global Management and BlackRock, to create a platform that will help customers build out AI infrastructure. The financial firms will raise $500 billion, or about 80 trillion yen, from outside investors to create a fund to buy Nvidia semiconductors.
Emerging AI companies will lease the chips from this fund. Nvidia will guarantee a certain portion of the residual value of the equipment after the lease ends, allowing AI companies to borrow at low rates backed by Nvidia's credit. On the 17th, it said it would guarantee up to $105 billion in lease payments for OpenAI's massive data center in Ohio in the US Midwest. By taking on OpenAI's default risk, Nvidia is supporting financing for a unit of SoftBank Group, which is developing the facility.
Rejecting the circular investment label
Chief Executive Jensen Huang stresses that these deals are not circular investments. But the structure does not change simply because Nvidia is not the source of the capital: it is still taking on customers' financial risk to encourage product purchases.
Nvidia is supporting customers through direct investment and credit because many emerging AI companies cannot raise the money needed to buy semiconductors on their own. Unlike hyperscalers such as Amazon.com, they are still in the process of monetizing and lack the credit strength to attract construction funding. Even the big tech companies are repeatedly raising capital through equity issuance and bond sales, limiting their ability to keep buying.
To broaden its customer base, Nvidia is effectively functioning as a fund for the AI industry and sustaining high growth. This pattern is not unique to Nvidia. In July, Advanced Micro Devices said it would directly invest up to $5 billion in Anthropic, while Broadcom is also supporting customer purchases of custom chips through what amounts to debt guarantees.
In 2025, Google signed at least four agreements with startups and others to guarantee repayment of financing related to AI data centers. The aim appears to be expanding sales of its own AI chips and securing computing resources. Alphabet's credit support for outside companies had reached $51.4 billion at the end of June.
In a July report, Moody's warned that the mutual dependence among AI developers, hyperscalers and chipmakers could become a circular structure that obscures genuine demand. There are no signs yet of a slowdown in AI demand, but the web of corporate ties centered on Nvidia is pulling in Wall Street and becoming more complex. If defaults emerge among startups, concerns are rising that the problem could become systemic. The AI capital market continues to expand as it attracts participants and new money.
Enjoyed this article? Share it with your network!