CoreWeave: Expansion still requires 24 billion–37 billion in financing; Morgan Stanley cuts price target
Execution capabilities remain recognized, but widening credit spreads make the cost and availability of expansion funding harder to assess. Morgan Stanley cut its price target from 99 dollars to 70 dollars and maintained its Equal-weight rating; the key is whether customer contracts can continue to support attractive financing terms.
Original source：Morgan StanleyFirst published on Unhidea：Original source date：
Key takeaway
Returns on CoreWeave’s projects remain attractive, but converting contracted capacity into operational capacity requires continued access to substantial financing at a reasonable cost.
Execution remains recognized, so why cut the price target to 70 dollars?
Credit spreads on CoreWeave’s unsecured bonds have widened by an average of approximately 300 basis points since the end of August 2026, and its shares have fallen more than 30% over the past month. Morgan Stanley continues to recognize its deployment speed and scale advantages, but changes in financing costs and availability have reduced visibility into the path for further expansion.
The bank cut its price target from 99 dollars to 70 dollars and maintained its Equal-weight rating. It lowered its base-case valuation range from 80–131 dollars to 50–115 dollars, shifting the valuation from roughly in line with peers to a modest discount, reflecting the constraints that financing changes place on future growth.
How much expansion can existing facilities support, and how much more is needed?
As of August 2026, the company had 4.2 gigawatts of contracted capacity. Morgan Stanley believes existing financing arrangements are better able to support adding approximately 350 megawatts of operational capacity within the year, reaching 1.85 gigawatts; bringing more contracted capacity into operation will still require additional funding.
Using capital expenditure of 37 billion dollars in 2026, operational capacity of 1.85 gigawatts at year-end and 3.3 gigawatts at the end of 2027 as its modeling starting points, the bank estimates that approximately 24 billion–37 billion dollars in additional debt financing will be needed after accounting for existing financing arrangements.
The range depends on the proportion of capital expenditure funded by debt and how much of the 2026 spending has already been used for 2027 capacity. The model assumes existing facilities are fully drawn before year-end; 2 billion dollars was available under the revolving credit facility in the second quarter, of which 1.2 billion dollars was drawn in August. The facility cannot be counted again as additional cash.
Why do the identity of customers and contract length affect borrowing terms?
Delayed-draw term loans (DDTLs) have been a principal financing channel. Morgan Stanley believes projects backed by contracts with large cloud providers of higher credit quality are more likely to secure competitive rates; as the customer base broadens and contract terms shorten, financing terms become less certain.
There is a trade-off: the bank lists more short-term contracts and higher GPU monetization prices as conditions for becoming more bullish, but short-term contracts also increase the variability of financing outcomes. Improving compute pricing and maintaining low-cost funding therefore need to be considered together.
Can project returns themselves support expansion?
Assuming deployment of Vera Rubin, a five-year contract, a revenue parameter of 15 million dollars per megawatt, no prepayment and no monetization after the contract ends, Morgan Stanley calculates an unlevered after-tax internal rate of return of 17.9% and a payback period of approximately two years.
Changes in pricing can materially alter returns: with no prepayment, reducing the revenue parameter to 12 million dollars per megawatt lowers the return to 9.3%. Prepayments can improve modeled returns, but they are also usually accompanied by lower pricing, so the two benefits cannot simply be added together.
Why do customer prepayments make cash flow and debt appear better?
Prepayments have characteristics of both compute purchases and financing, yet they are recorded in operating cash flow and listed as deferred revenue on the balance sheet. As a result, compared with financing through external borrowing, similar projects show higher operating cash flow and lower reported debt in their early stages.
The company must also recognize non-cash interest on prepayments, add it to deferred revenue and recognize it as revenue over the contract term, resulting in cumulative recognized revenue exceeding the cash ultimately received. Morgan Stanley therefore models future funding needs using an external-financing approach, so that cash flow and net debt reflect the financing burden of expansion.
