CoreWeave Inc shares fell 6.4% to $65.25 on Friday morning, extending a weekly loss to nearly 22% as credit markets priced in heightened default risk. The stock slid further in premarket trading, losing a third of its value over the last seven sessions.

Despite the sell-off, Bank of America reiterated its Buy rating and $140 price target, about 105% above current levels. BNP Paribas analyst Daniel Wang also maintained a positive outlook with a $169 price forecast, implying 157% upside from the October 14 close of $65.80. The divergence highlights the tension between credit market skepticism and equity analyst optimism regarding the company's growth trajectory.

Credit markets signal distress

Credit markets are sending a far less hopeful signal than equity bulls. CoreWeave’s five-year credit default swaps hit a record 1,194.5 basis points on Thursday, according to ICE Data Services data reported by Bloomberg. That price implies a more than 60% chance of default within five years, by standard market calculations.

The yield on CoreWeave’s 9.625% bond due 2032 rose above 15.4% on Thursday, with the bond trading near 78 cents on the dollar, a level widely treated as the line where a borrower is considered financially strained. The yield on its 9.75% bond due 2031 stood above 15.3%. S&P Global rates CoreWeave B+, which is in junk territory. The selloff coincided with a hotter U.S. inflation report that pushed traders toward betting on a December Fed rate hike, raising refinancing costs for heavily indebted companies.

Analysts flag distinct risks

Bank of America analyst Tal Liani said Wednesday’s drop “likely reflects concerns over financing hurdles for AI infrastructure players.” His team raised its forecasts, expecting revenue growth of 105% in 2027 and 63% in 2028, up from 102% and 58% previously.

Two levers drive the upgrade:

• Capacity: CoreWeave had 1.5 gigawatts of active power at the end of the second quarter, with expectations of 4.8 gigawatts by the end of 2028.
• Pricing: Annual revenue per watt is projected to rise from about $10 today to $14 by the end of 2028, as Nvidia Corp.’s next-generation Vera Rubin chips roll out.

The $140 target values CoreWeave at 26 times the bank’s estimate of 2027 operating profit, measured as enterprise value to EBIT. Cloud data center peers trade at 14 to 19 times.

BNP Paribas analyst Daniel Wang urged investors to distinguish between two types of debt: AssetCo financing, which supports infrastructure assets, and ParentCo borrowing, which funds broader corporate needs. Wang sees limited repayment risk associated with AssetCo debt because cash flows from the underlying infrastructure should comfortably cover operating costs, interest and principal payments. However, ParentCo debt presents a greater challenge, with implied yields on certain unsecured obligations reaching 15%, making traditional corporate borrowing increasingly expensive.

Cash needs remain substantial

Bank of America’s model shows why bondholders are nervous. It forecasts negative free cash flow of about $67 billion combined from 2026 to 2028. To fill that gap, it assumes about $59 billion of new borrowing.

Interest expense nearly doubles over the period, from $3.08 billion in 2026 to $5.9 billion in 2028. Operating profit is not expected to cover interest payments fully until 2028, the same year the bank sees CoreWeave’s first full-year profit. In the second quarter, CoreWeave generated $679 million in operating cash flow while spending about $6.4 billion on equipment. Net interest expense more than doubled to $640 million.

Wang noted that CoreWeave may need to rely on convertible debt or equity issuance to finance expansion until credit investors gain confidence in its ability to fund growth internally. He expects the company's financial position to improve as more graphics processing units become operational and newer contracts with better economics generate revenue.

Sector-wide pressure

Ludovic Subran, chief investment officer at German insurer Allianz, described the situation as a “Darwin moment” for data centre bonds, where the weakest credits struggle to survive. Rival neocloud Nebius Group N.V. is down about 20% this week.

Bank of America’s Michael Hartnett sees the pressure moving up the food chain. He cited 3% inflation-adjusted yields, the Nov. 3 midterms and weak trading in AI “spenders and lenders.” The spenders are CoreWeave’s customers: Amazon, Alphabet, Meta, Microsoft and Oracle. Their AI spending is expected to reach $807 billion this year. As a result, their combined free cash flow is forecast to swing from $191 billion in 2025 to minus $63 billion this year, according to consensus estimates compiled by Bank of America.

What the numbers show

The divergence between the bond market and analyst theses hinges on timing and debt structure. The banks bet that contracted revenue arrives on schedule and grows into the debt, while bond traders question what happens if financing becomes more expensive first. With interest expenses doubling by 2028 and operating profit not covering them until then, the company relies heavily on continued access to capital markets to bridge the gap between heavy upfront investment and eventual profitability. Wang specifically highlighted that evidence that CoreWeave can meet its corporate debt obligations without repeatedly raising capital could restore investor confidence.

Technology research firm Gartner projects that 80% of neocloud providers will cease independent operations by 2030. This consolidation is driven by a significant shortfall in enterprise and consumer AI revenue compared to rapid capital spending on infrastructure.

The warning highlights risks for retail-popular sector stocks including CoreWeave Inc. , Nebius Group NV, and IREN Ltd. Gartner estimates that more than $3.5 trillion in AI infrastructure spending could be exposed to impairment by the end of 2030. In 2026, AI providers are projected to spend nearly $1 trillion on infrastructure, with an additional $6 trillion forecast between 2027 and 2030.

Revenue Shortfall and Impairment Risk

Will Sommer, senior director analyst at Gartner, stated that through 2030, AI data center owners would need to generate a cumulative $8.5 trillion in revenue to cover announced capital spending, financing costs, and a 12% return on invested capital. However, Gartner forecasts total attributable enterprise and consumer AI revenue will reach no more than $3.6 trillion during this period.

Sommer noted that because most AI-native providers use asset-light models, less than half of AI revenue will be earned by providers that own significant AI infrastructure. This imbalance creates a structural challenge for companies relying solely on AI-optimized Infrastructure-as-a-Service (IaaS).

The Neocloud Squeeze and Consolidation

While major hyperscalers possess massive, diversified revenue sources to sustain investments for years, neoclouds must earn returns strictly on AI-optimized IaaS and associated platforms to sustain operations. Gartner estimates that AI-optimized IaaS spending will grow from $66 billion in 2027 to $174 billion in 2030.

Sommer stated this growth will be too low for most AI compute providers to recover their AI investments before the current generation of AI semiconductors is replaced. Consequently, cost pressure and a lack of a differentiating business model will force over 80% of neocloud providers to cease independent operations by the end of the decade.

A continued trajectory of infrastructure investment without sustainable returns could leave more than $4 trillion in asset value exposed to write-down risk through 2032, which Gartner notes is the earliest the 12% return rate might be achieved.

What the Numbers Show

The data reveals a stark divergence between capital expenditure requirements and projected revenue generation. While infrastructure spending is forecast to exceed $7 trillion (combining 2026 and 2027-2030 figures), total attributable AI revenue is capped at $3.6 trillion. This implies that even if all AI revenue were captured by infrastructure owners, it would cover less than half of the required outlay, let alone the additional $4.9 trillion needed to achieve the specified return on invested capital. This gap underscores why asset-light models and diversified hyperscalers are better positioned than pure-play neoclouds.

Market Performance of Key Stocks

Recent market performance reflects growing investor caution regarding these valuations.

Note: CRWV performance since listing in March 2025; NBIS since listing in October 2024.