CoreWeave and Nebius Each Lose Roughly Half Their Market Value in One Month
The simultaneous selloff across neocloud stocks suggests investors are repricing the entire AI infrastructure lending model, not just individual company performance.
CoreWeave and Nebius, two of the highest-profile artificial intelligence cloud infrastructure companies trading on US exchanges, have each lost approximately 50 percent of their market value over the past 30 days, according to reporting by CNBC published November 9, 2026. Both companies are scheduled to release earnings reports this week, adding further scrutiny to their financial positions.
The declines arrive as a broader category of companies known as neoclouds faces intensifying investor concern over three overlapping risks: weakening near-term AI infrastructure demand, high levels of debt used to finance GPU clusters, and regulatory uncertainty. Bloomberg reported on November 9, 2026 that investors are scrutinizing CoreWeave, Nscale, and other cloud providers over their financial stability.
CoreWeave, which completed its initial public offering on the Nasdaq in March 2025 and is headquartered in Roseland, New Jersey, built its business model around leasing Nvidia GPU capacity to AI developers. The company financed large portions of its data center buildout through debt secured against its hardware assets, a structure that critics have labeled a potential "doom cycle," according to the CNBC report. In that scenario, falling demand from AI customers reduces revenue, which in turn makes debt service more difficult, which then constrains the ability to invest in new capacity.
The mechanics of the risk are specific to neoclouds as a business type. Unlike hyperscalers such as Amazon Web Services, Microsoft Azure, or Google Cloud, which carry cloud infrastructure as one revenue line among many, neoclouds derive substantially all of their revenue from renting AI compute capacity. That concentration means revenue is more sensitive to changes in AI workload demand.
CoreWeave's largest disclosed customer has been Microsoft, which signed a multi-year agreement for GPU capacity. The existence and terms of that agreement were disclosed in CoreWeave's S-1 registration statement filed with the Securities and Exchange Commission prior to its IPO. Any revision or non-renewal of large customer contracts would be a material event requiring SEC disclosure; no such filing has been reported as of November 9, 2026.
Nebius, which trades on the Nasdaq under the ticker NBIS, was reconstituted from assets of the Russian technology company Yandex following a corporate restructuring completed in 2024. Its operations are now headquartered in Amsterdam but its shares trade on a US exchange, making it directly relevant to US investors. The company's approximately 50 percent decline over the past month mirrors CoreWeave's move, according to the CNBC report, despite the two companies having different customer bases and geographic footprints.
The parallel declines in both stocks suggest that the market is applying a category-level discount to neocloud business models rather than reacting to company-specific news. Whether that discount reflects permanent revaluation or temporary sentiment will depend in part on what earnings reports from CoreWeave and Nebius reveal about contract backlog, revenue concentration, debt covenants, and cash burn rates. Those figures are not yet public as of this reporting.
Debt structure is a central point of concern. CoreWeave disclosed in its IPO-related SEC filings that it carried billions of dollars in debt secured against its Nvidia hardware. The value of that collateral is tied to the market price and usable lifespan of GPU hardware, both of which can change as newer chip generations arrive. If hardware values decline faster than the debt is retired, the collateral cushion for lenders shrinks.
The neocloud sector grew rapidly during 2023 and 2024 as demand for AI training compute exceeded what hyperscalers could immediately supply. Neoclouds filled that gap by moving quickly to acquire Nvidia H100 and H200 GPUs and lease them to AI startups and enterprises. That window of excess demand appears to have narrowed as hyperscalers expanded their own capacity and as some AI developers reduced or redistributed their compute spending.
What remains unknown as of November 9, 2026, is the specific revenue and contract data that earnings releases will provide. Key disclosures to watch include the percentage of revenue from the top one or two customers, the remaining term on existing compute contracts, capital expenditure plans for 2027, and any amendments to existing credit facilities. Those figures, once released in earnings reports or accompanying SEC filings, will allow analysts to assess whether the selloff reflects accurate credit risk or an overcorrection.