Oppenheimer reiterated its "Outperform" rating on AI cloud computing service provider CoreWeave, Inc. (CRWV) on October 27 and maintained its $150 price target. Although CoreWeave's shares have fallen 37% over the past month and its unsecured bonds have also been sold off due to rising interest rates and concerns over GPU equipment depreciation, Oppenheimer believes the bond market may be overestimating the company's financial risk.
Through asset liquidation and cash flow stress tests, the firm pointed out that even under extreme scenarios, CoreWeave may still have sufficient assets to repay its unsecured debt, and the actual useful life of existing GPU equipment may be longer than the accounting depreciation period. Based on the closing price of $56.43 on October 26, the $150 target price implies potential upside of about 166%.
$83 Billion in Recoverable Assets Cover Debt, Oppenheimer Says Bond Selloff Is Overdone
Oppenheimer analyst Param Singh noted that the recent selloff in CoreWeave's unsecured bonds mainly reflects market concerns about rising interest rates, the company's debt scale, and the possibility that GPU equipment could quickly become obsolete. To assess the company's actual debt repayment capacity, Oppenheimer conducted an asset liquidation analysis.
According to the firm's estimates, CoreWeave has about $83 billion in recoverable assets, corresponding to about $42 billion in pro forma debt and about $37 billion in other liabilities. Under this model, total recoverable assets exceed the combined total of those two liability categories by about $4 billion. Singh believes that even in an extreme liquidation scenario, unsecured creditors could still recover the full principal. However, this conclusion relies on Oppenheimer's estimates of asset recoverable value and does not mean the company would necessarily be able to sell equipment at that price in an actual liquidation.
Oppenheimer also believes the economic useful life of GPU servers may be underestimated by the bond market. Singh cited a CoreWeave contract for NVIDIA A100 GPU services running through 2029, noting that the related servers could achieve a useful life of as long as nine years, well above the six-year accounting depreciation assumption. This means that even if GPU equipment is depreciated year by year in the financial statements, the equipment may continue to generate revenue as long as customers remain willing to pay for the related computing services. Oppenheimer believes existing long-term customer contracts show that the commercial value of some older GPU models can last longer, so the market's worry about rapid equipment obsolescence may be too pessimistic.
Assuming Data Center Expansion Stops, Annual Free Cash Flow Could Reach $9 Billion to $16 Billion
In addition to the asset liquidation analysis, Oppenheimer also designed two hypothetical scenarios to assess CoreWeave's cash generation capacity if it stops expanding data center scale. The first scenario is that CoreWeave stops expanding commissioned power capacity in 2027, when the related scale is expected to reach 1.9 gigawatts. Oppenheimer estimates that after the business enters a stable operating phase, the company could generate about $9 billion in annual enterprise free cash flow (FCFF).
The second scenario assumes the company continues expanding through 2028 and stops adding new construction after commissioned power capacity reaches 3.3 gigawatts. At that scale, CoreWeave's stable-period annual enterprise free cash flow could reach about $16 billion.
Oppenheimer Hypothetical Scenarios: Stop Expansion in 2027 | Stop Expansion in 2028
Commissioned Power Capacity: 1.9 GW | 3.3 GW
Stable-Period Annual Enterprise Free Cash Flow: About $9 Billion | About $16 Billion
Enterprise free cash flow is the cash flow available to all capital providers of a business before paying creditor interest and distributing cash to shareholders, and it is not equivalent to cash that can be directly used for share buybacks or dividends. Singh emphasized that these two no-expansion cases are only stress tests and are not Oppenheimer's forecast of CoreWeave's future operating strategy. The firm still expects the company to continue executing its original data center expansion plans. These assumptions are mainly intended to illustrate that if CoreWeave slows capital expenditures in the future and converts more already-deployed GPU resources into recurring revenue, its cash flow performance could improve significantly.
Bullish on CoreWeave Software Advantages, Expects Q3 Beat and Raised Full-Year Guidance
Regarding the long-term competitive landscape, Oppenheimer believes CoreWeave is positioned to become an important winner in the AI infrastructure industry, with a key advantage being that it not only provides GPU computing resources but has also built a differentiated software service system. Singh noted that in the future, the focus of competition among AI cloud computing companies will gradually shift from simply expanding the number of GPUs to providing complete software platforms that meet enterprise customers' model inference and other workload needs. For data center operators, how to increase revenue per megawatt of power and reduce the operating cost of computing tasks will become important factors affecting investment returns.
Oppenheimer believes that compared with cloud service providers that mainly rent out GPU computing power, CoreWeave's software capabilities can help improve customer usage efficiency, strengthen service differentiation, and improve long-term profitability. On short-term performance, Oppenheimer expects CoreWeave's third-quarter financial report to beat market expectations and for the company to slightly raise its full-year 2026 revenue guidance, bringing full-year revenue growth to about 160% year over year. This judgment is mainly based on newly built data center computing power gradually entering operation and support from computing service pricing.
Looking ahead to 2027, Singh expects CoreWeave's revenue may receive further support as more data centers are put into use, the share of software service revenue increases, and higher-value enterprise and AI inference workloads grow. Overall, the core basis for Oppenheimer's maintenance of the $150 target price this time is that the market may be underestimating the long-term revenue capacity of CoreWeave's existing GPU assets and overestimating its debt risk. However, the firm's optimistic judgment still depends on continued equipment utilization, execution of customer contracts, and the smooth commissioning of new data centers. Whether actual cash flow can be delivered will be the key to subsequent valuation recovery.
Comments