이미지: CNBC Technology · METAL LAB 편집
Summary
- Nvidia CEO Jensen Huang appeared on CNBC's "Mad Money" to rebut criticism that the company's financial support for AI firms is a way to artificially inflate revenue
- Nvidia has guaranteed $105 billion for an Ohio data center campus and struck a financing partnership worth up to $500 billion with major Wall Street firms for data center construction
- Announced the same day, fiscal 2027 second-quarter revenue came in at $96.2 billion, more than double the year-ago figure, with data center revenue up 117% to $89 billion
On the same day Nvidia reported $96.2 billion in fiscal 2027 second-quarter revenue, CEO Jensen Huang sat down with Jim Cramer on CNBC's "Mad Money" and directly defended the company's growing financial support for other AI firms. The interview came as criticism mounted over Nvidia playing both investor and financial guarantor to a wide range of companies — from model developers like OpenAI and Anthropic to the neoclouds that lease out Nvidia chips. Asked whether this support was really just a way to artificially inflate revenue, Huang framed it differently: today's AI startups are an unprecedented first generation that needs tens of billions of dollars just to get started, then tens of billions more to reach profitability.
Why the circular-financing criticism is growing
Nvidia has notably expanded the scale of its financial backing in recent weeks. According to disclosures our newsroom confirmed on August 17, Nvidia partnered with SB Energy to secure land, power, and buildings at a site in Portsmouth, Ohio, naming OpenAI as the tenant — and in the CNBC interview, it emerged that Nvidia guaranteed $105 billion for construction of that campus. On top of that, the company recently announced a partnership with major Wall Street firms to raise up to $500 billion for data center construction. And on August 24, reports surfaced that Nvidia was in talks to invest in Perplexity at a valuation north of $30 billion.
The core criticism centers on what's known as "circular financing" — a company lending money to a customer, who then uses that same money to buy the company's own products. Critics worry this can inflate revenue in ways disconnected from real demand. The comparison that keeps coming up is the early-2000s telecom equipment industry, where vendors used similar arrangements to boost sales before the dot-com bust wiped them out.
Huang's counterargument
Huang argued that frontier AI labs simply don't yet have the financial track record or credit history to qualify for investment-grade backing on their own, which is exactly why they need a partner like Nvidia. He said Nvidia wants to be both an equity investor and a partner to these companies, betting that if they grow their businesses on Nvidia's ecosystem, Nvidia grows right along with them. Pressed on whether Nvidia would be left holding the risk if any single customer stumbled, he said the company's computing infrastructure can be redeployed to other customers and other workloads, which limits its exposure to any one investment. "I think the risk is low," Huang said flatly.
The numbers behind his confidence
The same day brought figures that back up Huang's confidence. Nvidia's second-quarter revenue more than doubled year over year to $96.2 billion, while data center revenue jumped 117% to $89 billion. The company also guided for roughly 70% revenue growth in fiscal 2028. Nvidia shares rose about 4% in after-hours trading following the report — though the stock is up just 12% for the year so far, a sign that investor unease about the broader AI trade hasn't gone away.
| Item | Scale | Note |
|---|---|---|
| Ohio data center campus guarantee | $105 billion | OpenAI slated as tenant |
| Wall Street financing partnership | Up to $500 billion | For data center construction |
| Perplexity investment talks | Valuation above $30 billion | Reported August 24 |
Editor's take
This isn't the first time Nvidia has moved from selling picks and shovels to lending customers the money to buy them. What's changed is the scale. A few years ago, this meant taking a small equity stake in a startup's funding round. Now it means guaranteeing $105 billion for a single Ohio campus and opening a $500 billion financing channel with Wall Street. Carrying that kind of exposure requires serious cash flow behind it — and that's exactly the case Huang is making with a $96.2 billion quarter.
Anyone who's watched the semiconductor industry for a while has seen this movie before. Intel once blended equity stakes into its foundry customer deals, and telecom equipment makers used to lend carriers money to buy their gear. But there's one key difference between then and now: back then, the circular financing structure was exposed once revenue growth slowed. Right now, Nvidia's growth rate is still close to triple digits. As long as the results keep coming in, the circular-financing critique stays a theoretical risk — but the moment growth slows, the same structure could easily be read in reverse.
For companies watching from here, two things stand out. First, any company trying to secure Nvidia chips may no longer be a pure buyer — it could end up becoming Nvidia's financial partner too. Second, as these large-scale guarantee and financing structures multiply, they're building a body of precedent that domestic data center and AI infrastructure investment decisions may increasingly draw on. Unless Nvidia's data center revenue growth stalls over the coming quarters, the circular-financing debate will likely keep making headlines — without actually slowing down Nvidia's investment strategy.
Comments