CoreWeave and Nebius earnings preview – can results halt the neocloud rout?

It has not, by any reasonable measure, been a good month to own a neocloud, with CoreWeave shares down 49% and Nebius off 48% as investors concluded that building data centres on borrowed money looks considerably less clever when borrowing costs are rising and the bond market has started asking awkward questions. Both companies now report third-quarter (Q3) results within days of each other – CoreWeave after the US close on 10 November and Nebius before the open on 12 November – and the timing could hardly be more pointed.

The awkward part for the bears is that, operationally, neither company has given the market much to complain about. Both beat revenue forecasts in each of the past two quarters, analysts still expect revenue to roughly double at CoreWeave and more than triple at Nebius in 2027, and mean price targets sit at well over twice current share prices. The problem is not the top line.

It’s what has been borrowed to produce it.

Why CoreWeave and Nebius shares have halved in a month

The sell-off has had less to do with anything either company has said than with a broader shift in attitudes towards AI-related debt, with Meta’s credit default swap (CDS) spreads hitting a record in late October after it raised its capex forecast yet again, and investors responding by selling the most leveraged end of the AI infrastructure trade – the neoclouds, which are spending several times their annual revenue on data centres. If even the hyperscalers are being asked to justify their spending, companies with a fraction of their balance sheets were always likely to face sharper questions.

CoreWeave has taken the brunt of it. The company carries roughly $50bn of debt, and in mid-October its 2031 bonds yielded almost 16% while CDS spreads widened to around 1200bp – levels the credit market normally reserves for borrowers it suspects may struggle to repay. Nebius, with a smaller debt load and a freshly signed $3.5bn financing facility, has nonetheless fallen almost as far, which suggests that investors have not been in a particularly discriminating mood.

What the Street expects from CoreWeave and Nebius

For CoreWeave, consensus looks for Q3 revenue of $3.55bn, with the StarMine SmartEstimate at $3.54bn – essentially in line, and implying a fractional miss of 0.06%. That fits with a beat that narrowed to 0.59% in Q2 from 5.73% in Q1, a sign that analysts may finally have caught up with the growth rate. EBITDA is forecast at $2.05bn, an implied margin of around 58%, though the company is still expected to post a loss of $1.16 per share, narrowing to $0.55 in Q4.

Nebius is forecast to report revenue of $908m, with the SmartEstimate slightly higher at $916m for a predicted positive surprise of 0.96%. Its Q1 and Q2 beats came in at 7.43% and 1.67% respectively, so the same narrowing pattern is visible here too. EBITDA of $353m implies a margin of about 39%, while losses are expected at $0.69 per share and could widen to $0.95 in Q4 before improving next year.

The bull case: why the sell-off may have gone too far

On the numbers alone, the case for both stocks remains almost embarrassingly strong. Consensus has CoreWeave’s revenue rising to $26.47bn in FY2027 from $12.86bn this year, growth of around 106%, while Nebius is expected to jump to $11.98bn from $3.37bn, a rise of roughly 255%, and both have contracted work to back those forecasts up – a revenue backlog of more than $129bn at CoreWeave and over $40bn of long-term contracts at Nebius.

Valuations have also come down to earth with some speed. CoreWeave trades on a trailing EV/EBITDA multiple of 14.6x, hardly the stuff of bubbles for a business forecast to double its revenue, while Nebius’s multiple drops to 6.4x on a forward basis from 191.4x on a trailing one. Mean price targets of $131.41 for CoreWeave and $286.53 for Nebius imply upside of around 183% and 121% respectively – although analysts, as is customary, have been rather slower to cut their targets than the market has been to cut the share prices.

Then there’s the newsflow. Palantir has named Nebius a preferred sovereign AI infrastructure partner, which could open up government revenue, while the October acquisition of Inferize should strengthen its inference offering. CoreWeave, meanwhile, is expanding into Denmark and building out a global AI network backbone reaching the UAE, India, Indonesia, Singapore, Australia, Korea and Japan from Q4.

The bear case: why debt may matter more than revenue

The difficulty is that growth funded by debt is only as attractive as the terms on which the debt can be rolled over. CoreWeave raised its capex forecast to $35-39bn in August, roughly three times its expected FY2026 revenue, and with bond yields near distressed levels the question is less whether customers want the capacity than whether lenders will keep paying for it. Any hint that the capex range is being trimmed for lack of funding, rather than by choice, could be read badly.

Nebius’s financing looks healthier, but not without its quirks. The new $3.5bn senior secured facility with Citi carries a 364-day maturity at SOFR plus 2.75-3.50%, which amounts to short-dated money funding long-lived assets, and management’s plan to use its contract book as the base for asset-backed financing will need to come off in a market that has suddenly become rather less generous.

Elsewhere, customer concentration remains a live issue for CoreWeave, given its exposure to Microsoft, OpenAI and Meta, and CEO Michael Intrator’s sales of nearly $1bn of stock since the IPO – mostly under prearranged trading plans – may surface as a governance question at an inconvenient moment. Nebius, for its part, may field questions about its former parent Yandex, which withdrew its 2026 forecast after attacks on data centres in Russia, even though the two are now separate businesses.

Why the earnings calls could decide the debate

In a normal quarter, a modest revenue beat and steady guidance would be enough. This is not a normal quarter. With both companies expected to come in close to consensus, the headline numbers may matter less than what management says about financing conditions, the timing of free cash flow and the breadth of the customer base, and investors will be looking to CoreWeave for some kind of free cash flow roadmap stretching out to 2030.

Margins also carry more weight than usual. CoreWeave’s implied 58% EBITDA margin is central to the argument that it can service its debt, while Nebius’s 39% is improving quickly from a lower base, and any slippage at either company could reinforce the idea that the economics of renting out GPUs are less attractive than the growth rates suggest.

What could signal a turning point

Several things could help persuade the market that the worst of the selling is done:

• CoreWeave confirming its $35-39bn capex range with funding clearly identified, rather than quietly trimming it.
• Credible timeline for positive free cash flow at CoreWeave, ideally with targets running through 2030.
• Net new bookings that broaden CoreWeave’s customer base beyond its biggest hyperscale and AI lab clients.
• Detail from Nebius on asset-backed financing against its $40bn contract book, and on how the 364-day facility will be refinanced.
• Evidence of inference workload growth at Nebius following the Inferize deal.
• A narrowing in CoreWeave’s bond yields and CDS spreads in the days after results, which may say more than the share price reaction.

If both companies can show that the money is lined up and the margins are holding, the market may decide that a 40%-plus fall priced in rather more pessimism than the numbers justify. If the calls are instead dominated by refinancing questions and vague answers, the halving of the share prices could prove to be a fair assessment rather than an overreaction.

CoreWeave & Nebius stock price – technical analysis

The CoreWeave share price – down 35% year-to-date – has been trading in a clearly defined downtrend channel since the beginning of October and trades at levels last seen in May 2025.

While no bullish reversal takes the CoreWeave share price above its late October $64.39 high, downside pressure is expected to retain the upper hand.

A fall through last week’s $40.58 low would open the way for the late April 2025 trough at $39.50 to be reached. Further potential downside targets in this scenario would be the March 2025 low at $36.00 and the April 2025 all-time low at $33.52.

CoreWeave daily candlestick chart

CoreWeave weekly candlestick chart

Were a bullish reversal to take the CoreWeave share price to above its 29 October $64.39 high – a possibility as the current November low hasn’t been confirmed by a lower reading of the Relative Strength Index (RSI) which is instead showing positive divergence on the daily chart – an advance towards the October high and 200-day simple moving average (SMA) at $90.17-to-$93.21 may ensue.

The Nebius share price – up around 55% year-to-date despite having fallen close to 50% from its early October peak – looks close to testing its January high at $110.50. It is part of a large support zone which consists of the mid-November-to-February sideways trading range between $110.50 and $73.52. This we would expect to hold, were it to be reached.

Nebius daily candlestick chart

Nebius weekly candlestick chart

Given that the Nebius share price has fallen over the past five consecutive weeks and positive divergence can be spotted on the daily RSI, a bullish reversal may soon ensue. For it to gain traction a rise and daily chart close above the 200-day simple moving average (SMA) and 28 October high at $178.38-to-$182.74 would need to take place. If so, the July-to-early August highs at $229.47-to-$230.30 may be revisited and perhaps the September-to-October highs at $254.74-to-$255.10.