Rebecca Bellan
Senior reporter at TechCrunch covering business, policy and emerging trends in AI. Her work has also appeared in Forbes, Bloomberg, The Atlantic and The Daily B...
“If you look at hyperscalers, Meta, Google, Amazon, they're building data centers, they're spending a lot of money, they're going into negative free cash flow, but they still have other businesses that will help them float. This is the business for these companies. So they're taking out stupid amounts of debt to fund a business where there are so many things that can go wrong.”
Equity · Startups & Venture · November 2026
Context
The Equity hosts were discussing whether neoclouds like CoreWeave and Nebius are good businesses after this week's earnings. Bellan explains why they carry more risk than the big tech companies building the same data centers.
Transcript
I think that's going to be one of the biggest challenges for AI-related companies in the coming weeks with their earnings is how much of this is tied to the uncertainty around NVIDIA versus actually what it says about their companies.
…overall, like, you know, Core Weave is down pretty significantly for the year. And my sense is not that, oh, like suddenly the business for NeoClouds changed very dramatically. It's that Wall Street suddenly did not necessarily feel as optimistic about the long-term prospects of these businesses.
Yeah, I mean, it's, there's a few moving parts that make NeoClouds very risky. And that has really affected their stock price since they went public, you know, specifically Core Weave and Nebius. And I think one of the big ones, like if you look at hyperscalers, right? You look at Meta, Google, Amazon, they're building data centers, sure. They're spending a lot of money. They're going into negative free cash flow for this, but they still have other businesses that will help them float. Like this is the business. For these companies. So they're taking out stupid amounts of debt to fund a business that is really, there's so many things that can go wrong, right? Like there's so many things that can go wrong when you're trying to do a build out. Just look at Nebius with their New Jersey data center. They're trying, well, they're not building it, but they're relying on a third party to build it. And there's all these issues related to that one in terms of the types of generators they were using that were, you know, not Koch, definitely didn't have permits for them. That's slowing things down. They're having to change the energy source. Like, so little things can…
and maybe you have some thoughts on this too, Rebecca, is does this undermine the theory that picks and shovels are what make money in a gold rush, right? I think a lot of people thought that there was going to be money in these tools, data centers, power, et cetera, surrounding the AI boom as…
software itself. And it's sounding like that's maybe not the case. I mean, yeah, I think it's hard to say. Like, how much does it cost to literally make a literal pick and shovel? Probably not that much. How much does it cost to make a data center? A whole lot more. And so much more can go wrong.
Yeah, I mean, it's something that I can imagine if everything works out great, then yeah, maybe that, you know, these numbers are going to be incredible in a couple of years. But like you said, there's a lot of uncertainty. Although I would also point out that something you said earlier, Rebecca…
More from Equity
