After the stock and bond selloff, Goldman Sachs Group, Inc. and JP Morgan came out in support of CoreWeave (CRWV.US): Strong returns on AI infrastructure projects, and the new cloud leader has been unfairly beaten down by the market.

Title context: After the stock and bond selloff, Goldman Sachs Group, Inc. and JP Morgan came out in support of CoreWeave (CRWV.US): Strong returns on AI infrastructure projects, and the new cloud leader has been unfairly beaten down by the market. Text: The current financial market debateunprecedented in its intensityover CoreWeave (CRWV.US), the stock at the very center of the "AI compute selloff storm," the "AI bubble burst narrative," and the NeoCloud selloff, is focusing on one question: Can the cash flow returns generated by AI compute assets in the future cover higher capital costs and support financing needs during the expansion phase? On October 19, Goldman Sachs Group, Inc. upgraded its rating from "Neutral" to "Buy," while cutting its target price from $139 to $106, which still implies about 72.7% upside from Friday's close of $61.39; MoffettNathanson maintained its "Neutral" rating and lowered its target price from $107 to $89, arguing that the company is not on the brink of collapse, but its massive financial obligations mean it still relies on continued capital market funding. Generally speaking, a rating upgrade and a target price cut are not contradictoryfor example, Wall Street financial giant Goldman Sachs Group, Inc. is now accepting the valuation discount caused by a deteriorating financing environment, while believing the share price decline has already been larger. Another Wall Street financial giant, JP Morgan, is bullish on CoreWeave's core because it believes the market has overinterpreted the funding gap caused by rapid capacity expansion as a lack of returns inherent in the business model. The institution's equity strategy team recently reiterated its "Overweight" rating on CoreWeave, emphasizing that data center demand has continued to strengthen in recent months, and that CoreWeave has locked in customers in advance through long-term "take-or-pay" contracts, reducing the vacancy risk of building first and finding demand later. Its competitive advantage also lies in integrating power, GPUs, liquid cooling, and high-speed networking into a truly deliverable computing cluster: announced planned capacity in the industry does not equal effective supply that can come online on time. For customers eager to deploy models, delivery speed and cluster reliability have commercial value and also support CoreWeave's pricing power. This contrarian rating upgrade or reiteration of bullish target prices by Wall Street institutions such as Goldman Sachs Group, Inc. came after a deep correction: CoreWeave fell 25.21% last week and was down 29.53% cumulatively from October 1 to October 16; the Philadelphia Semiconductor Index recorded eight consecutive declines over the same period, plunging about 11.4% last week and down about 24% from its recent one-year high. South Korean memory chip giants SK Hynix and Samsung Electronics also recently suffered concentrated selling, while the KOSPI rebounded 1.76% on October 19. Financing pressure is also reshaping primary market pricing: Nscale, backed by NVIDIA Corporation, is seeking an IPO valuation of about $30 billion, but some investors are only willing to accept $16 billion$19 billion, equivalent to demanding a discount of about 37%47%a price-cutting demand close to halving the valuation, and not a completed financing valuation. Capital markets are raising their requirements for delivery certainty, cash recovery, and return on capital. The reason CoreWeave's bond metrics have recently become a new barometer for AI trades reflects the market's beginning to re-examine: Can strong compute demand be converted into cash flow sufficient to support expansion and debt servicing? The deep contradiction currently in focus is that GPUs, liquid cooling, high-speed optical interconnect network infrastructure, and power connections require upfront investment, while customer payments are realized gradually as clusters are delivered and services are performed; the faster the expansion, the larger the gap between cash flow from mature projects and new investment. When credit spreads widen sharply, even if U.S. Treasury yields fall, corporate financing costs may not decline in tandem. Financing pressure may therefore be transmitted through project delays, slower equipment procurement, and changes in delivery pace to GPUs, storage, optical interconnect, and even the entire semiconductor supply chain, depressing the market's pricing of future orders and profit growth. Investment-grade bonds remain relatively stable while lower-rated AI debt comes under pressure, meaning capital is more strictly distinguishing between companies "able to finance growth" and companies that "must keep financing in order to grow"which also explains why, as credit markets flash warnings, Goldman Sachs Group, Inc. and JP Morgan remain bullish on CoreWeave: long-term pre-sale contracts increase revenue visibility, cluster delivery capability supports order fulfillment, and renewals of older GPUs extend the cash recovery cycle; JP Morgan indicates that the unlevered returns of projects in its model remain above its assumed cost of capital. The bull-bear debate thus centers on: Can projects with long-term returns obtain sufficient funding to smoothly bridge the gap between construction and collection? New contract pricing, cash return per megawatt, and actual financing conditions are therefore more decisive for re-rating potential than total order value alone. If new financing is successfully completed, credit spreads stabilize, and deliveries continue to convert into cash, the stock market rebound can gain more solid support; financing divergence may also drive orders toward operators with stronger financial strength and higher delivery efficiency, causing AI compute industry expansion and corporate valuation divergence to occur simultaneously. At the institutional level, JP Morgan recently reiterated its "Overweight" rating, believing that data center demand has continued to strengthen in recent months and that CoreWeave's project economics still support healthy profits and cash returns. Among 35 analysts aggregated by MarketBeat as of October 19, 25 rated it Buy, 7 Hold, and 3 Sell, with an average 12-month target price of $138.82, 126.1% above Friday's close; however, it should be noted that this average includes target prices from different dates and cannot be regarded as all institutions having simultaneously completed their revaluation after this round of credit shock. MoffettNathanson strikes a cautious stance: "Not on the brink of collapse," but financing dependence remains It is reported that analysts from well-known Wall Street investment institution MoffettNathanson said CoreWeave is "not on the brink of collapse," but warned that its financial obligations are "staggering in scale." The MoffettNathanson warning came just after CoreWeave's stock suffered a brutal week. Surging borrowing costs raised questions about whether its debt-financed AI expansion is sustainable. MoffettNathanson believes CoreWeave is not "on the brink of collapse." However, the brokerage on Monday still cut its target price on the AI cloud infrastructure provider from $107 to $89 and maintained its "Neutral" rating. According to The Fly, the institution said CoreWeave still has multiple financing options, although some may not be ideal. However, it warned that the company has already taken on "staggering financial obligations," and ultimately it still depends on whether capital markets are willing to fund its business. CoreWeave shares plunged as debt concerns unsettled global tech investors. The latest assessment came after CoreWeave suffered a brutal week: from October 9 to October 16, it fell more than 25% cumulatively. On Wednesday of last week, the selloff intensified further. Higher-than-expected U.S. inflation data reinforced expectations of further rate hikes, triggering market concerns about financing costs for highly indebted AI infrastructure companies. CoreWeave's borrowing costs became a particularly prominent focus. The yield on its U.S. dollar bonds maturing in October 2031 briefly rose above 16% on Thursday, while the cost of debt default insurance exceeded 1,000 basis points. Notably, despite having $104 billion in backlog, CoreWeave still carries $35.6 billion in debt. CoreWeave disclosed that as of June 30, its total debt was $35.6 billion. Second-quarter net interest expense rose to $640 million from $267 million a year earlier, highlighting the cost of financing expansion. According to an internal memo seen by media, CEO Michael Intrator responded to the concerns, saying the market sometimes "misunderstands and acts irrationally." As of the end of June, the company had about $104 billion in contracted revenue backlog. How do other analysts view CoreWeave? Despite turbulence in the bond market, some analysts believe the selloff has exaggerated CoreWeave's financial risk. It is reported that Goldman Sachs Group, Inc. on Monday local time upgraded CoreWeave from "Neutral" to "Buy," and although it cut its target price from $139 to $106, the new target still implies about 73% potential upside. Goldman Sachs Group, Inc. acknowledged that credit market volatility has raised financing concerns, but expects CoreWeave's annual recurring revenue per megawatt and unlevered return on investment to improve over the next 24 months, citing tight AI compute supply and the company's differentiated competitive advantages versus other cloud service providers. Yorkville Ives analyst Dan Ives said in a media interview last Friday that the market interpretation behind the selloff is a "false narrative." He believes strong AI demand and CoreWeave's backlog make this pullback a buying opportunity. JP Morgan analyst Samik Chatterjee likewise believes the market's concerns about rising bond yields and returns on AI infrastructure investment are "overdone," noting that data center demand has strengthened and CoreWeave's project economics remain solid. How do retail investors view CoreWeave? Despite the recent sharp drop in the share price, retail investor sentiment toward CoreWeave stock remains at "extremely bullish," and discussion volume is also at an extremely high level. The chart above shows CoreWeave retail sentiment as of 7 a.m. U.S. Eastern Time on October 19. Source: Stocktwits. Year to date, CoreWeave shares have fallen more than 22%. By comparison, the iShares Future AI and Tech ETF (ARTY.US), which also holds the stock, and the Invesco QQQ Trust Series ETFthe ETF that fully tracks the Nasdaq 100 (QQQ.US)have risen nearly 40% and 16%, respectively, over the same period. Why can't $100 billion in orders stop the credit alarm? Goldman Sachs Group, Inc. bets on improving returns per megawatt The transmission chain of this selloff is that inflation and rate shocks raise financing thresholds, and wider credit spreads then further compress shareholder value. The 10-year U.S. Treasury yield has now fallen back to about 5.08%, so it can no longer be described as a one-way surge; however, the yield on CoreWeave bonds maturing in 2031 briefly hit 16% last week, and the five-year credit default swap (CDS) spread exceeded 1,100 basis points. According to media reports, this corresponds to an implied cumulative five-year default probability of about 63%it is market pricing affected by recovery rate assumptions, risk premiums, and liquidity, not an objective measure of the company's actual default probability. The core market concern is the timing mismatch between upfront capital expenditure and later contract collections. CoreWeave's second-quarter net interest expense reached $640 million, up about 140% year over year; operating cash flow for the quarter was $679 million, while cash capital expenditure was $6.422 billion. Even with about $104 billion in contracted revenue backlog, revenue recognition still depends on delivery and service availability conditions, and order value cannot be directly equated with net cash available for debt servicing. Construction or power delays postpone billing, but may not simultaneously postpone interest, rent, and equipment payments, so strong demand and tight financing can absolutely coexist. Goldman Sachs Group, Inc.'s bullish call is precisely built on a re-estimation of this relationship: the bank expects upside in CoreWeave's annual recurring revenue per megawatt (ARR/MW) and unlevered internal rate of return (Unlevered IRR) over the next 24 months. At the same time, Goldman Sachs Group, Inc. expects the blended cost of debt to rise by 100200 basis points and emphasizes that the company mainly uses asset-backed financing, so the 13.5%15% worst-case yield on unsecured bonds cannot be directly applied to all new financing. Its core judgment is that the market may be overestimating further deterioration in financing costs while underestimating improvement in returns on compute assets. From a technical and commercial mechanism perspective, this judgment needs to be reflected in cluster efficiency: high-speed networking and storage reduce GPU waiting, task scheduling improves resource utilization, and fault identification and recovery reduce ineffective computing, thereby lowering the cost for customers to complete equivalent training or inference tasks. CoreWeave provides a dedicated platform covering these areas, and its value lies in increasing effective compute output and delivery reliability, supporting pricing and renewals. Only if revenue growth per megawatt, after deducting new equipment, power, and O&M inputs, still improves project returns will technical advantages debt-servicing capacity and shareholder returnsthis is also the key to whether this potential rebound in semiconductors and broad tech stocks can be sustained: financing channels remain open while signed orders convert into collectible cash on schedule. JP Morgan's bullish case rests further on contractual cash returns. According to its illustrative calculations, the unlevered internal rate of return on a typical five-year contract is in the "mid-teens," above the modeled 9%11% blended cost of capital; after repaying asset-level debt, it can also contribute residual cash equivalent to about 16% of total contract value to the parent company. However, this 16% is not a shareholder net profit margin, as the parent company still has to bear its own financing and operating obligations. JP Morgan believes tight compute supply gives new contracts an opportunity to incorporate higher financing costs into pricing, thereby preserving project return headroom. The key is whether new projects can continue to earn cash returns above the cost of capital, rather than whether the market yield on a single existing bond spikes; the latter affects future financing but does not mean all debt immediately bears the same high interest rate. The continued monetization of older GPUs adds further room to this return model. The company disclosed that A100-related contracts have been extended to 2029, indicating that earlier-generation equipment can still serve adapted workloads such as data processing and model fine-tuning; after the initial contract covers equipment investment and repays related debt, renewal revenue may still contribute cash after deducting power, leasing, and maintenance costs. At the same time, JP Morgan attributes current cash consumption mainly to overly rapid investment in new capacity: mature projects collect cash year by year, while new projects require large upfront construction and equipment payments, and its model estimates that only when mature capacity reaches about 45 times annual new capacity will operating cash earnings be more capable of covering expansion spending. This is a scale-dependent judgment relying on continuity of delivery, pricing, and financing, not an automatic improvement in cash flow. Therefore, the positive judgments of JP Morgan and Goldman Sachs Group, Inc. jointly point to thisCoreWeave may have project value obscured by credit panic, but re-rating requires contract value to actual debt-servicing capacity. Goldman Sachs Group, Inc.'s upgrade to "Buy" focuses on its competitive position and future growth financing capacity; the massive financial obligations emphasized by MoffettNathanson remind the market that long-term project profitability and short-term corporate liquidity are still two different things. What can truly drive a sustained rebound is the simultaneous realization of three things: projects coming online on schedule, improving cash returns per megawatt, and stabilizing new financing costs.