CIFC

Are neocloud stocks a buy after this week's sell-off?

Seeking Alpha analysts Danil Sereda, Petri Dish Reports, and Carla Magliocco weigh in.

Danil Sereda: In a recent article, I compared Nebius and CoreWeave to find the future outperformer, and following the sell-off that we're seeing now in both of them, I'm not changing my opinion that Nebius looks superior in the long run.

First, it's the management team. Nebius has a CEO who built Yandex, the biggest tech conglomerate in the CIS region, and one of the biggest in Eurasia in general. The team is very experienced, and Nebius holds a massive advantage over CoreWeave in terms of labor costs and flexibility.

Second, their balance sheet remains quite strong. I know it's going to take on incremental debt in the next few years, but I like the way management structures the new debt loads and attracts money through convertibles or direct stakes from Nvidia and others.

If we recall why neocloud stocks are dropping so much, we'll find that it was due to Gartner predicting that ~80% of neoclouds will evaporate by 2030. Nebius is unlikely to be one of those evaporated companies, in my opinion. Not in a today-like environment of rising demand for AI workloads and all that.

Coming back to the main question, I'd like to answer that not all neclouds deserve a Buy rating on the dip. Nebius deserves it for the reasons I listed above. CoreWeave doesn't, because its debt levels look extreme, and the odds of evaporating are mathematically higher. At least that's how I see the setup.

Ex-miners such as IREN , Cipher Digital , or Wolfspeed are also not Buys because they seem to lack the superior tech filling that Nebius and CoreWeave have. They're mainly just selling GPU compute, where the asset base is at risk of becoming obsolete. Neclouds need to have SaaS-filling and tech features to give customers more offerings and make sure their workloads don't break at some point. Small ex-miners don't seem to be capable of doing that, in my opinion.

In short, I'd only go with Nebius on this dip and ignore all other tickers.

(Danil Sereda leads the SA investing group Beyond the Wall Investing.)

Petri Dish Reports: The Gartner report had a big impact on market sentiment. My main argument is that we should treat AI infrastructure demand and neocloud shareholder returns as two completely different investment questions. The reality is that the industry is going to expand pretty rapidly, while some infrastructure providers struggle with financing costs, GPU depreciation, or even weak margins and dilution.

When we compare big players like CoreWeave , Nebius , Iren , and Applied Digital , we need to focus on three areas.

First, who captures the economics? We have to compare their contract pricing, margins, and differentiation rather than simply relying on a player's backlog figure or projected revenue.

Second, who can actually finance their growth sustainably? Debt obligations, cash flow, and construction costs are going to have an impact.

Lastly, who owns assets that actually retain value? What I mean by this is that we need to distinguish GPU-dependent businesses from companies whose economics are tied to land, power, and some hefty long-term data center leases.

Everyone realizes that choosing which company will win is the hard part here, but first we should understand that treating all neoclouds as one investment category is going to overlook some very meaningful differences in their business models and financing risks.

Carla Magliocco: The market is going through a panic spiral in AI similar to the one we saw in the last week of July. And where there's panic, buy! I took the opportunity to expand my position in Nebius .

It's true that this time the economic factors are more deteriorated than in July, especially because taking on debt for big investments has become more expensive. All equities are suffering in this context, and AI is suffering as well. But the cost of debt hits neoclouds such as Nebius , CoreWeave , Hut 8 , and IREN harder, because their growth models depend on large expenditures by hyperscalers. Remember, however, that this context calls into question the financing capacity, not the development nor the profitability of AI.

That’s why now is the time to stay calm and put any available liquidity, big or small, to work. And those who have no more liquidity, just hold your position. Neoclouds offer a key and highly specialized service for AI hardware: computing power. If you have doubts about this thesis, where the expected returns in this bottleneck are huge, selling wouldn’t be the best alternative.

• Trending AI Stocks