NVIDIA invests 100 billion in OpenAI, a real gamble or a capital performance?

NVIDIA and OpenAI announced a blockbuster collaboration worth $100 billion, which is less of a conventional business deal and more of a meticulously planned capital performance.

According to the agreement announced on Monday local time, NVIDIA will provide OpenAI with the advanced chips urgently needed for building, training, and hosting AI models (such as the recently released GPT-5) in large-scale data centers.

However, an unexpected twist in the agreement is that NVIDIA will also inject $100 billion into OpenAI in phases to purchase its unlisted shares. This amount far exceeds the total funds raised by OpenAI over the past decade. According to Crunchbase data, OpenAI has raised a total of $7.2 billion since its establishment ten years ago.

Analysts believe that this collaboration between the two companies is not just a business partnership but a capital strategic performance. This transaction avoids NVIDIA's financial risks by using equity investment rather than traditional supplier financing. For OpenAI, this massive funding supports its grand narrative of general artificial intelligence; for NVIDIA, this move intensifies competition in the AI field, prompting other companies to seek similar collaborations and solidifying its market dominance.

How is this investment different from the telecom hype?

At first glance, this agreement resembles the supplier financing model common during the telecom boom in the early 2000s. Companies like Nortel, Lucent, and Motorola provided funding to customers to maintain rapid revenue growth, but after the bubble burst, they bore a significant amount of bad debt.

Although the AI boom has somewhat overheated, this situation is a bit different. NVIDIA is participating in OpenAI through equity investment, meaning OpenAI does not have to repay NVIDIA like a loan. This investment, while risky—equity may depreciate, or OpenAI's future orders may not materialize—poses almost no existential pressure for NVIDIA. After all, NVIDIA has about $100 billion in free cash flow each year and a market capitalization of $4.5 trillion, and founder Jensen Huang is fully capable of undertaking such an investment.

A meticulously designed "performance"?

The real mystery of this transaction lies in its necessity. Currently, NVIDIA is almost the only choice in the market for the large-scale supply of advanced AI chips, although other companies, including OpenAI, are also developing alternatives. Therefore, it seems that Jensen Huang has no need to secure a customer through a massive investment.

For OpenAI, however, it does need funding. It is estimated that the company's revenue this year is expected to reach $12 billion, which is far from sufficient to support its multi-billion dollar capital expenditure plans. But as a recognized leader in the AI field, OpenAI should have no problem raising tens of billions of dollars in financing through the market in batches If this deal looks like a "performance," perhaps that is its very significance. Both parties can benefit from the impression that "the AI race is entering a higher level."

For OpenAI, securing a commitment of hundreds of billions in funding strongly supports its grand narrative of advancing towards artificial general intelligence that surpasses human intelligence. Announcing plans to build more large-scale data centers will make its valuation of up to $500 billion more convincing to investors and keep this "vision flywheel" turning.

For NVIDIA, the AI field will witness a "life-and-death showdown" between OpenAI and its peers, a notion that is highly beneficial to its business interests. Companies like Google, Meta, and privately held Anthropic are racing towards "superintelligence," firmly believing that those who underinvest will be eliminated by history. Now, these competitors have even more reason to seek similar massive deals with NVIDIA