NVIDIA's Neocloud Deals Follow a Financing Playbook That Has Failed Before

NVIDIA spent much of the past year insisting it’s nothing like Enron. Tech critic Ed Zitron’s latest piece argues the more useful comparison is the vendor financing schemes that preceded the collapse of telecom giants Lucent and Nortel.

What Zitron said

Writing on Where’s Your Ed At, Zitron argues NVIDIA has built a financing structure that pushes risk onto everyone except itself.

“NVIDIA CEO Jensen Huang has employed many of the same tactics used by Lucent, Nortel, and many of the big dot-com busts, but has been smart enough to make everybody else carry the risk,” he writes. Unlike Lucent’s approach, “direct vendor financing like Lucent did with Winstar (where it effectively loaned its customers money to pay it with),” Zitron says NVIDIA works through “neoclouds”, companies like CoreWeave, Nebius, and IREN that exist mainly to buy NVIDIA’s chips and rent out the computing power to other businesses.

NVIDIA operates as their “early stage investor, IPO anchor, post-IPO investor, $6.3 billion customer and data center lease backstop.” In plain terms: it invests in these companies early, commits to buying shares when they go public, buys a large chunk of their capacity itself, and guarantees their data center leases. That combination, Zitron says, has let neoclouds raise “tens of billions of dollars’ worth of debt” from outside investors. He sums up what NVIDIA has become as an “asset management and marketing firm that happens to sell semiconductors,” and notes it now represents more than 7% of the total value of the NASDAQ stock exchange.

Who Zitron is

Zitron is the CEO of the media relations firm EZPR and writes Where’s Your Ed At, a newsletter focused on tech industry criticism. He also hosts the Better Offline podcast.

What he gets right, and where it’s overstated

NVIDIA’s dual role as both investor in, and dominant customer of, the neocloud companies it partly funds is a real and unusual structure, not something Zitron is inventing. Financing that lets a supplier’s own customers borrow money that ultimately buys that supplier’s product, even indirectly through outside investors rather than a direct loan, is the kind of arrangement that made Lucent’s and Nortel’s collapses worse: when the customer can’t pay, the demand that justified the original investment disappears too.

The publicly available excerpt does not establish that NVIDIA’s neocloud partners face the financial distress that preceded Winstar’s failure. The Lucent and Nortel comparison remains Zitron’s framing, not a demonstrated outcome.

Why it’s notable

Most coverage of AI infrastructure spending leads with dollar figures: how much NVIDIA is selling, how large the data centers are. Zitron’s point is structural instead: who is holding the risk when the investor funding a company’s growth is being repaid with revenue from the same chips it financed.

What it means for builders

If your product depends on GPU capacity from a neocloud rather than a hyperscaler (a large, general-purpose cloud provider) like AWS, Azure, or Google Cloud, know that some of these providers’ finances are more tied to NVIDIA’s own investment decisions than a typical vendor relationship would be. That’s not a reason to switch providers today. It means the pricing and availability you’re planning around could be more exposed to NVIDIA’s capital strategy than to straightforward supply and demand for chips.

The practical step is the one that applies to any infrastructure dependency: understand who actually owns and finances the hardware behind your GPU bill, not just who sends the invoice.

End of article