Investors are questioning neoclouds' business model but Nebius says AI computing prices are rising

BERLIN — Nebius is seeing AI computing prices rise more than expected, a top executive told CNBC on Wednesday, offering a bullish view as neocloud firms' business model comes under increasing scrutiny.

Investors have grown wary after CoreWeave, one of the biggest neocloud players, saw its cost of borrowing spike. The CoreWeave developments and questions around the amount of debt being used to fund the data center buildout have weighed heavily on other neoclouds like Iren and Nebius, whose stocks have dropped sharply over the last month.

Marc Boroditsky, chief revenue officer at Nebius, said his company wasn't in CoreWeave's position because "we're not carrying the kind of debt that they are."

He also said that if the AI infrastructure company brought its 2027 capacity online, it would sell out. There could be "the same type of dynamic" in 2028, supporting a positive longer-term demand outlook, he said, as companies race to secure access to computing power.

When asked whether he sees the price per megawatt continuing to rise ahead of expectations, Boroditsky replied: "Yeah."

Nebius is among a new wave of Neocloud companies that are investing billions to buy up Nvidia's graphics processing units, build data centers and rent that capacity.

Nebius ran an auction for computing capacity on Tuesday, which showed continued price rises, the revenue chief said. "We're seeing continued upward mobility in terms of the kind of price per megawatt," he told CNBC.

Price per megawatt, a key measure among new cloud companies, shows how much a company can charge customers for a certain amount of computing capacity.

Chase Lochmiller, CEO of Crusoe, an Nvidia-backed AI company that is building data centers, was also bullish on pricing, saying the company was seeing rising prices for GPUs, including older Nvidia chips.

"We're seeing higher pricing today for A100s than we saw in 2020 or 2021. We're seeing higher pricing for H100s than what we saw when they were first deployed in 2023," Lochmiller said.

"There's tremendous demand and a lot of it is because the underlying model capabilities have gotten so much more advanced that people are just creating so much more value out of the compute infrastructure."

The comments come amid an investor debate over how quickly Nvidia's GPUs lose their economic value as more powerful chips enter the market.

The ability to continue renting out older chips at attractive prices is crucial to AI infrastructure companies that need to recover their investments and generate profits, while spending billions on GPUs.

Lochmiller's comments suggest that, at least for Crusoe, older generations of Nvidia chips are commanding higher rental prices than when they were first deployed, potentially strengthening the economics of its AI cloud business.