On October 27, Oppenheimer reiterated its 'outperform the market' rating for AI, a cloud computing service provider known as CoreWeave ( CRWV ), and maintained a target price of $150 per share. Despite the company's stock price falling by 37% over the past month, and its unsecured bonds being sold off due to rising interest rates and concerns about the depreciation of GPU equipment, Oppenheimer believes that the bond market may have overestimated the company's financial risks. The firm's asset liquidation and cash flow stress tests indicate that even in extreme circumstances, CoreWeave may still have sufficient assets to repay its unsecured debts, and the actual useful life of the existing GPU equipment could be longer than the accounting depreciation period. Based on the closing price of $56.43 on October 26, a target price of $150 represents a potential upside of approximately 166%.

I. $83 billion in recoverable assets to cover debts; AuburnHai assumes excessive bond selling

Oberhaimer analyst Param Singh (Param Singh) pointed out that the recent selling of CoreWeave unsecured bonds mainly reflects market concerns about rising interest rates, corporate debt levels, and the possible rapid obsolescence of GPU equipment.

To assess the company's actual debt-paying ability, Oppenheimer conducted an asset liquidation analysis. According to the bank's estimates, CoreWeave owns approximately $83 billion in recoverable assets, which correspond to about $42 billion in contingent liabilities and around $37 billion in other debts.

According to this model, the total value of recoverable assets is about $4 billion higher than the combined total of the aforementioned two types of liabilities. Singer believes that even in an extreme liquidation scenario, unsecured creditors may still be able to recoup the full principal amount.

However, this conclusion relies on Oppenheimer's estimation of the recoverable value of the assets, and it does not mean that the company will necessarily be able to sell the equipment at that price during actual liquidation.

Oberheimer also believes that the economic lifespan of the GPU servers may be underestimated by the bond market. Taking a NVIDIA A100 GPU service contract that lasts until 2029 as an example, Singer pointed out that the relevant servers could achieve a usage cycle of up to nine years, which significantly exceeds the six-year accounting depreciation assumption.

This means that even though the GPU equipment is depreciated over the years in the financial statements, as long as customers are still willing to pay for the related computing services, the equipment may continue to generate revenue.

AuburnHai assumes that existing long-term customer contracts indicate that the commercial value of some older models GPU can be maintained for a longer period, therefore concerns about the rapid obsolescence of these devices may be overly pessimistic.

II. Assuming the expansion of data centers is halted, it could generate between $9 billion and $16 billion in free cash flow each year.

In addition to asset liquidation analysis, Oppenheimer also designed two hypothetical scenarios to assess the cash generation capability of CoreWeave after ceasing to expand the data center scale.

The first scenario is that CoreWeave will cease to expand the already operational power capacity in 2027, by which time the relevant scale is expected to reach 1.9 gigawatts. Oppenheimer estimates that after the business enters a stable operation phase, the company may generate approximately $9 billion in annual free cash flow ( FCFF ).

The second scenario assumes that the company continues to expand until 2028, and stops adding new capacity after the existing operational power capacity reaches 3.3 gigawatts. At this scale, the stable annual free cash flow of CoreWeave could reach approximately 16 billion US dollars.

Enterprise free cash flow is the cash flow available for use by the enterprise's capital providers before paying interest to creditors and distributing cash to shareholders, and it is not equivalent to the cash that can be directly used for stock repurchases or dividends.

Singer emphasized that these two scenarios of halting expansion are only for stress testing purposes and are not predictions by Oppenheimer regarding the future business strategy of CoreWeave. The bank still expects the company to continue with its original data center expansion plan.

These assumptions are mainly used to illustrate that if CoreWeave slows down its capital expenditures in the future and converts more of the already utilized GPU resources into ongoing revenue, its cash flow performance could significantly improve.

III. Optimistic about the advantages of CoreWeave software; expects third-quarter performance to exceed expectations and raises full-year guidance

In terms of the long-term competitive landscape, Obenhai assumes that CoreWeave is likely to become an important winner in the infrastructure industry of AI. The key advantage lies in its ability to not only provide GPU computing resources but also to have established a differentiated software service system.

Singer pointed out that in the future, the focus of competition among AI cloud computing companies will gradually shift from simply increasing the number of GPU to providing comprehensive software platforms that meet the model inference and other workload needs of corporate clients.

For data center operators, how to increase revenue per megawatt of power and reduce the operating costs of computing tasks will become important factors affecting the return on investment.

AuburnHai assumes that, compared to the main cloud service providers that rent out GPU computing power, the software capabilities of CoreWeave help to improve customer efficiency, enhance service differentiation, and improve long-term profitability.

In terms of short-term performance, Oppenheimer expects that the CoreWeave third-quarter financial report is likely to exceed market expectations, and has slightly raised its revenue guidance for the entire year of 2026, aiming for a year-on-year revenue increase of about 160%.

This judgment is mainly based on the gradual deployment of computing power in the newly established data centers, as well as the support for the pricing of computing power services.

Looking ahead to 2027, Singer expects that with more data centers coming into operation, an increasing proportion of revenue from software services, and a growth in high-value enterprise and AI inference workloads, the revenue of CoreWeave may see further momentum.

Overall, the main rationale for Oppenheimer to maintain a target price of $150 this time is that the market may be underestimating the long-term revenue potential of CoreWeave's existing GPU assets and overestimating its debt risks. However, the bank's optimistic assessment still relies on the continuous utilization of equipment, the execution of customer contracts, and the successful operation of new data centers. Whether the actual cash flows can be realized will be key to subsequent valuation corrections.