Nvidia's role as the AI central bank raises bubble concerns

The AI kingmaker's portfolio

Nvidia has moved beyond simply selling hardware to become a primary investor in the artificial intelligence landscape. The company recently committed $3 billion to Meta's share sale, adding to a massive portfolio of stakes, loans, and guarantees. This strategy has positioned Nvidia as what some analysts call the central bank of the AI boom.

The company has taken direct equity in leading AI developers including OpenAI, Anthropic, and Elon Musk's xAI. Its reach extends globally, with investments in French developer Mistral, British self-driving startup Wayve, and a partnership with South Korea's SK Group. These deals ensure that the most prominent names in the field are tied directly to Nvidia's success.

Beyond equity, Nvidia acts as a customer of last resort by providing financial backstops. For example, it committed to buying up to $6.3 billion of unused computing capacity from CoreWeave if no other customers take it. These moves help keep demand running high for the processors that power the world's vast data centers.

Why Nvidia is funding the field

Experts suggest that Nvidia's massive profits leave it with few options other than investing back into the infrastructure that uses its chips. Michael Cusumano, a professor at the MIT Sloan School of Management, noted that channeling profits into equipment-tied investments is a smart move. It allows the company to support the ecosystem without competing directly with its biggest customers.

By investing in the infrastructure, Nvidia avoids the potential antitrust complaints that would arise if it ran its own data centers. Instead, it fills the gaps in the market where other entities might not be investing enough. Gil Luria of D.A. Davidson noted that while this approach is unprecedented, it appears to be working to maintain momentum.

CEO Jensen Huang has defended this strategy by highlighting the unique needs of frontier AI labs. He argues that these startups require tens of billions of dollars to get started rather than just millions. Nvidia's investments effectively provide these companies with the necessary time and capital to succeed in a high-cost environment.

Concerns over circular financing

Critics are beginning to draw comparisons between these deals and the dot-com era of the late 1990s. During that period, companies sometimes inflated their growth by investing in startups that then used that money to buy the parent company's products. This creates a cycle where it becomes difficult to determine if there is actual, sustainable demand for the technology.

While there is nothing inherently wrong with a hardware provider helping its clients, the proximity of these relationships is a point of contention. Paulo Carvao of the Harvard Kennedy School suggests that the closer these ties become, the harder it is to assess true demand. If the money is simply moving in a circle to keep the lights on, the underlying business model may be fragile.

Luria argues that not all demand is created equal in this ecosystem. He distinguishes between a company like Microsoft buying chips for an enterprise customer and a neocloud company buying chips simply because Nvidia gave them the money to do so. The latter represents a higher risk of artificial inflation compared to demand driven by end-user utility.

Concentration and contagion risks

A major concern for Wall Street is the high level of concentration within Nvidia's revenue streams. According to the company's annual report, just two direct customers accounted for 22 percent and 14 percent of its revenue in the 2026 fiscal year. This heavy reliance on a few entities makes the company vulnerable to any shifts in those specific clients' spending.

The web of overlapping investments also increases the risk of financial contagion if a weaker partner fails. If a company like CoreWeave runs into trouble, the fallout could hit chipmakers, data-center builders, and financiers all at once. These concerns were highlighted this month when shares of several Nvidia-linked neocloud companies fell at least 30 percent in October.

The market for AI chips remains limited by supply, which currently keeps demand high. Carvao estimates that demand will continue to grow for another 12 to 18 months before it begins to level off. However, if spending cuts occur due to a cooling market, the interconnected nature of Nvidia's bets could lead to a rapid decline across the board.

The long-term outlook

Despite the risks of a bubble, some analysts believe the demand for Nvidia's hardware is still in its early stages. Michael Cusumano expects high demand for these chips to continue for three to five years. This suggests that while the financing structure is complex, the underlying need for computing power remains significant for the foreseeable future.

Nvidia's strategy has successfully turned it into the world's most valuable company by market capitalization. By acting as a financier, it has secured its position as the kingmaker of the AI revolution. The company is essentially building the financial and physical foundation upon which the next generation of technology will sit.

The next few months will be critical as the market watches for signs of cooling credit for AI projects. Investors will be looking to see if the neocloud companies can find sustainable customers beyond Nvidia's backstops. The balance between smart infrastructure investment and artificial demand inflation will define Nvidia's trajectory in the coming years.