CRWV: The customer cash arrives before the shareholder cash

The 60-second read

Five things to know

1. Operating improvement is real. Revenue grew sequentially by $1.086B and GAAP operating profit turned positive; adjusted margin also improved from Q2.
2. Customer funding is not free owner cash. A $2.127B deferred-revenue inflow exceeded Q3 operating cash flow of $1.975B. This is a financing-quality diagnostic, not a restatement of cash flow.
3. Interest still dominates. Q3 net interest was $846M versus $379M adjusted operating income, and Q4 interest guidance is $1.05B–$1.15B.
4. The buildout requires continued outside capital. Nine-month CFO less cash capex was negative $17.920B; customer deposits, borrowings and equity issuance can finance growth but carry obligations.
5. The evidence has limits. Q3 statements are unaudited. The latest quarterly filing reviewed, for Q2, reported material control weaknesses; the Q3 release supplied no controls assessment.

Three scenarios, one probability-weighted number

November 11, 2027 prices, distinct from present fair values. Zero dividend assumed; probabilities are author judgments.

Signposts: what would change our mind

Guidance-derived ranges and author decision rules are different. Controls are a qualitative hard checkpoint; the date of the next filing is not invented. A cheaper quote alone does not satisfy our BUY conditions.

Price, security and expectations

The reference quote is $60.77 per Nasdaq Class A common share at the November 11, 2026 U.S. close. The November 11 close was up 24.7% from the November 10 close of $48.73. This is the next trading-day close versus the release-date close; the exact release time remains unconfirmed. That reaction shows a large change in market expectations; it does not prove that the company has become self-funding or establish which announcement caused the move. The saved price history begins after the March 2025 public listing, so pre-IPO earnings cannot be assigned a public-market reaction.

Our model uses 559 million shares, the Q3 basic weighted average, and therefore an equity capitalization of $33.970B. This is a valuation denominator, not an exact fully diluted or point-in-time share count. The provider reports about $33.517B of market capitalization using its own denominator. We keep the model consistent rather than mix the provider capitalization with a different per-share denominator. There is no ADS ratio or currency translation: the quoted security and accounts are both in U.S. dollars.

The gap between our $67 base value and the quote is modest beside the uncertainty. The price is 9.3% below fair value using fair value as the denominator; reaching fair value would be 10.3% upside using the current price as the denominator. Our downside-weighted present scenario value is $58.70, below the quote. This combination supports HOLD rather than an automatic BUY after a sharp decline from prior highs. Price history is context; future cash economics determine the investment case.

A stronger quarter with a funding condition

Revenue reached $3.661B, compared with $1.365B in the year-earlier quarter and $2.575B in Q2. Computed growth was 168.2% year over year and 42.2% sequentially. GAAP operating income improved to $207M from a $49M loss in Q2. This is a genuine improvement in the income statement after depreciation and ordinary operating expenses; it should not be dismissed merely because cash investment is large. The hard question is how much more capital is needed to sustain the improvement.

Adjusted operating income was $379M, or 10.4% of revenue using the unrounded calculation, compared with $128M in Q2. The issuer displays rounded margins of 10% and 5%. Year over year, however, adjusted margin declined from roughly 15.9% to 10.4%, even though adjusted profit rose. Sequential improvement and year-over-year margin compression can both be true. Adjusted EBITDA of $2.256B is useful for comparing operating scale before specified charges; it is not a measure of cash left for stockholders.

Management raised full-year revenue guidance to $13.0B–$13.3B and adjusted operating income guidance to $1.275B–$1.375B. Subtracting nine-month reported results implies Q4 revenue of $4.686B–$4.986B and adjusted operating income of $747M–$847M. These are computed implications, not separately issued Q4 guidance. At the midpoints, implied Q4 adjusted operating income of $797M remains $303M below the $1.100B midpoint of Q4 net-interest guidance, before taxes and any other below-the-line items.

Full-year capex guidance is $37B–$39B, with the lower bound raised from $35B. The company expects to deliver capacity sooner, so faster customer demand can pull investment forward. That creates a productive-growth possibility and a financing requirement at the same time. The annualized exit revenue target of $20B–$21B is the final month multiplied by twelve; it is neither full-year 2026 revenue nor a guaranteed 2027 result. The packet does not present a provider estimate mismatch as a verified earnings beat.

One segment, several kinds of exposure

CoreWeave reports one operating and reportable segment. Its chief operating decision maker uses consolidated net loss as the segment profit measure, with cost of revenue, technology and infrastructure, sales and marketing, and general administration as the significant expense categories. We therefore show the consolidated economics rather than invent separate training, inference or software segment profits. Product names and workload examples establish commercial breadth; they do not establish independently disclosed margins.

Active power exceeded 1.7 gigawatts at quarter-end, while contracted power exceeded 4.6 gigawatts. More than 1.3 gigawatts of contracted power were international. These are lower-bound operational disclosures, not exact point estimates, and active power is different from contracted power. A power agreement does not mean that equipment has been installed, customer workloads have started, or revenue has been collected. The gap between the measures is part of the future delivery obligation and capital requirement.

The reported $132.5B revenue backlog includes remaining performance obligations plus other estimated committed amounts. Recognition is conditional on delivery and service availability. It is valuable evidence of demand, but it is not cash, current revenue or equity value. The previous Charged Alpha Q2 episode centered on backlog duration; this quarter instead tests the financing chain from customer cash through infrastructure spending to the residual shareholder claim.

Customer concentration remains important even as the issuer adds names and workloads. The audited 2025 report attributed 67% of annual revenue to Microsoft. That is a dated annual observation, not an assertion that the same share persisted in Q3 2026. Customers can also be competitors or technology partners, and supplier concentration can expose the company to pricing, delivery and product-cycle risk. Diversification should be measured in recognized revenue, counterparty obligations and cash receipts, not simply the length of a customer list.

The bridge from operating improvement to loss

Q3 adjusted operating income of $379M reconciles to GAAP operating income of $207M after $160M of stock-based compensation and $12M of acquired-intangible amortization. Acquisition-related costs in that bridge were zero for the quarter. The operating result then absorbs $846M of net interest, adds $13M of other income and reaches a $626M pretax loss. A tax provision of $83M produces the $709M net loss. No minority allocation is needed to explain this bridge.

The year-earlier quarter benefited from a $127M tax benefit, while the current quarter incurred a tax expense despite a consolidated pretax loss. The swing matters when comparing net losses. A negative computed effective tax rate of approximately 13.3% does not mean the tax system pays CoreWeave a benefit; it is a positive expense divided by a negative pretax result. We do not capitalize that unusual ratio into the long-run valuation. The model instead uses a stated 25% normalized tax assumption.

Adjusted EBITDA removes $1.889B of depreciation and amortization, $846M of interest, $160M of stock compensation and the tax provision, with the other-income adjustment reversing that income. Depreciation can differ from near-term replacement cash, but equipment eventually needs replacement or becomes less competitive. Treating the whole depreciation add-back as a permanent distributable benefit would overstate owner economics. Conversely, current expansion capex is not all maintenance, so simply capitalizing the present cash deficit forever would understate the value of a successful buildout.

Nine-month GAAP operating income was only $14M on $8.314B of revenue, while net interest was $2.022B and the net loss was $2.075B. The latest quarter improved materially within that year-to-date picture. Our valuation requires the improvement to continue and the capital intensity to decline as the installed base matures. It also keeps stock compensation inside modeled operating costs. The share denominator and future financing terms remain separate risks; a non-cash charge can still transfer economic value.

Cash quality and reporting limits

Operating cash flow of $1.975B is positive, but the quarter included a $2.127B increase in deferred revenue in the cash-flow reconciliation. Subtracting only that inflow produces negative $152M. This diagnostic isolates the scale of one source of cash; it is not normalized operating cash flow and does not forecast what cash would have been without customer contracts. It leaves all other working-capital changes in place, including a $1.741B receivables use, and cannot be described as an accounting correction.

The latest saved Q2 2026 filing says disclosure controls were ineffective and material weaknesses remained unresolved, including deficiencies involving information technology, segregation of duties and personnel. The issuer said the deficiencies had not resulted in material misstatements at that reporting date. That statement is narrower than a remediation conclusion. The Q3 release is unaudited, and no Q3 Form 10-Q appeared in the saved current submissions at the source checkpoint. The next filing must be read for changed figures, new weaknesses or explicit remediation supported by testing.

Ending receivables of $4.280B divided by quarterly revenue and multiplied by 92 days produces a rough 107.6-day proxy. It uses ending rather than average receivables and a rapidly growing quarter, so it is not a clean collection-time estimate. We retain the formula and limitation instead of treating it as conventional days sales outstanding. Inventory is not a meaningful separately disclosed operating metric for this service model in the release. Missing inventory detail is not evidence of zero physical equipment investment.

Cash conversion ratios against negative net income are not useful rankings here. The more informative tests are customer collections, the deferred-revenue movement, cash capex, debt service and whether operating margins improve without accumulating receivables. The scorecard therefore distinguishes “not meaningful” from a favorable score. A clean audit opinion on annual financial statements and unresolved control weaknesses can coexist; neither label should be substituted for the other.

Debt, leases and the next dollar of funding

At September 30, unrestricted cash was $4.122B and marketable securities were $62M. We exclude $2.282B of restricted cash from freely available funds. Recourse and non-recourse debt together were $46.873B, and finance-lease liabilities were $649M. Subtracting unrestricted cash and marketable securities gives model net debt of $43.338B including finance leases. Project-level non-recourse financing is still an obligation of assets that support the consolidated business; it is not ignored when valuing that business.

Operating-lease liabilities were a further $18.697B. We disclose them separately and do not add them again to enterprise debt because the modeled operating profit already bears ordinary operating-lease expense. A valuation that capitalizes leases would need a matching adjustment to profit and cash flow. Mixing a lease-adjusted debt number with unadjusted operating costs can double-count the same obligation. The model’s explicit treatment matters more than adopting the largest possible debt headline.

Q3 cash spending on property, equipment and software was $9.441B, leaving CFO less cash capex at negative $7.466B. Across nine months, the same arithmetic was $5.638B less $23.558B, or negative $17.920B. This cash-flow measure is not identical to the company’s full-year capital-expenditure guidance, which can include different timing and non-cash additions. It also does not separate maintenance from growth investment. The funding gap is real; the return on that growth spending is the unresolved valuation question.

Management expects about $9B of customer prepayments and deposits over the next three months. Those funds can reduce the amount raised elsewhere, but create service commitments and are not dividends waiting to be paid. The quarter also included a roughly $4.2B convertible issue, a new $2.6B financing facility and authorization for an at-the-market program of up to 35 million Class A shares. Authorization is not evidence that all shares have been issued. Financing access is a strength until its price, collateral or dilution overwhelms the return on new capacity.

Three routes, one demanding operating path

Our base valuation is $67 per share, rounded from a weighted $67.08 across three routes. These are author assumptions, not company guidance or analyst consensus. The routes share an explicit operating path: revenue of $26B in 2027, $40B in 2028, $56B in 2029 and $70B in 2030, with EBIT margins rising from 8% to 19%. Revenue then reaches $116B and a 24% EBIT margin in 2035. This is a demanding growth and execution case, not a conservative extrapolation of the latest quarter.

The enterprise cash-flow route uses EBIT after a 25% tax charge, subtracts capex net of depreciation and amortization, and subtracts working-capital investment equal to 1% of incremental revenue. Net capex declines from $12B in 2027 to $3.5B in 2035. This explicitly charges for future investment rather than treating EBITDA as cash. It does not assume customer prepayments can grow forever to fund the entire business. Negative early free cash flows reduce present enterprise value; subtracting future debt again would double-count those funding needs.

At a 13% required enterprise return and 2.5% terminal growth, the DCF yields $57.23 per share after subtracting current net debt once. Terminal reinvestment is growth divided by a 20% assumed return on incremental capital: 12.5% of terminal after-tax operating profit. Approximately 78.8% of enterprise value comes from the discounted terminal value, so terminal economics dominate the answer. The annual model discounts 2027 as the first one-year bucket from the November 2026 valuation date; it does not pretend to be a precise daily cash forecast.

The second route applies twelve times modeled 2030 EBIT of $13.3B, discounts that enterprise value at 13%, includes the present value of the intervening positive and negative enterprise cash flows, and subtracts current net debt. It produces $76.45. The third route finances after-interest cash deficits with net debt at a 7% net-interest assumption. Current tax equals 25% of positive pretax income and is floored at zero in loss years: no cash refund or immediate use of loss carryforwards is assumed. Actual interest-deduction limits, loss utilization, jurisdictional taxes and payment timing can differ. It estimates 2030 equity earnings of $6.309B, applies twelve times earnings and discounts at a 15% equity return. It produces $77.44. That route contains interest and debt financing already, so no second enterprise-debt subtraction is made.

The current quote implies roughly a 24.8% terminal EBIT margin if the rest of the base DCF assumptions are held fixed. This reverse calculation is an interpretation of our model, not a market forecast. It shows how much sustained profitability is embedded even after the stock decline. We place 50% weight on the cash-flow route and 25% on each cross-check. A 30% margin-of-safety discount to the rounded base value is $46.90, approximately $47; financing conditions and reporting reliability must also remain acceptable.

Fresh targets, stale aggregates and our disagreement

The saved ratings snapshot is broadly positive: 23 Buy, 8 Hold and 2 Sell recommendations, with no Strong Buy or Strong Sell entries in that provider field. Its aggregate target consensus is $114.68, with a $45–$200 range. Those aggregates mix analysts and update dates. They should not be described as a clean post-Q3 consensus, and the target horizon of each entry is not independently established by the numerical feed.

The November 11 target-news snapshot includes Rosenblatt reiterating Buy with a $160 target, D.A. Davidson at $45, Stifel at $60, Cantor Fitzgerald at $100 and Mizuho at $68. Target-news entries establish dated reported targets; the title does not always establish the full rating, so we avoid inferring one from a dollar figure. The dispersion is economically plausible for a highly leveraged growth business whose value depends heavily on financing and terminal margins.

We disagree with the aggregate Buy stance because the present cash claim is more uncertain than the operating-growth headline. Our $67 is present fair value, while the $77.05 base scenario is a one-year price and the $67.51 weighted scenario is a probability-weighted one-year price. A Street target whose horizon is unspecified should not be compared with all three as if they were identical. Alignment is partial on demand and growth, but not on the margin of safety required for a new common-share purchase.

The provider annual-estimate feed is archived but is not accepted as a consistent adjusted-earnings consensus. Its 2026 EBITDA estimate of about $3.763B is already below the issuer’s nine-month adjusted EBITDA of $4.923B. This could reflect stale data or differing definitions. We show the issuer’s guidance and our assumptions directly rather than create a manufactured beat or a misleading consensus bridge from incompatible measures.

Execution and the financing bargain

Management delivered faster revenue growth and improved sequential operating profitability while raising guidance. Those are meaningful execution positives. Raising the capex floor at the same time shows the bargain: near-term delivery can require earlier cash spending. A successful capital allocator must earn an adequate return on that spending after financing, equipment replacement and customer commitments, not simply demonstrate the ability to raise more money.

The capital structure also affects governance. Class A shares carry one vote each, Class B carries ten and Class C carries no votes; the co-founders held all outstanding Class B shares at December 31, 2025. Outside shareholders have an economic claim without proportionate voting influence. Deloitte & Touche has audited the company since 2024. The annual audit context does not remove the unresolved internal-control weaknesses discussed above, and the latest quarter’s release is unaudited.

Our management questions are specific. How much of next year’s capex is necessary to maintain existing contracted economics rather than add capacity? What customer cash has been received, what remains conditional and what service credits or refund terms accompany it? How does the company match shorter-duration contracts to equipment and financing lives? What changes would cause an earlier useful-life revision? Finally, which tested controls will support an explicit remediation conclusion, and when will investors receive that evidence?

The strongest case on each side

The bull case is that contracted customer demand and specialized infrastructure create a durable growth platform. Faster deployment can bring revenue forward, spread operating costs and establish customer relationships that extend beyond the first equipment cycle. Customer prepayments can be an efficient financing source when service obligations are priced well. If financing costs decline and mature infrastructure generates attractive returns after replacement spending, the common equity can participate in substantial growth. Our bull value allows both better terminal margins and a lower required return.

The bear case is that today’s revenue and adjusted EBITDA growth require repeated investment whose financing cost and technological risk are underestimated. Customers can be concentrated, suppliers can have bargaining power, and equipment can lose economic usefulness faster than accounting estimates. Delayed power or deployment can postpone revenue while debt and lease obligations continue. A financing solution that protects operations can still dilute existing shareholders. The lower scenario is therefore not a floor: refinancing failure or adverse contractual terms can impair equity much more severely.

The base case is also vulnerable to modeling choices. We assume the business can reach $116B of annual revenue, sustain a 24% terminal EBIT margin and reduce net capital intensity. We assume access to debt funding and a stable 559M share denominator for comparability, while separately identifying dilution risk. These are consequential conditions. The model cannot establish that every future contract is profitable, that refinancing will be available, or that unresolved control weaknesses will be remediated. Very high uncertainty is an input to entry discipline, not decorative language.

Check the cash, then revisit the price

The next immediate checkpoint is the Q3 Form 10-Q when available. We will reconcile the release with the filing and read its controls, liquidity, customer concentration and debt disclosures. The source checkpoint did not locate that filing; we do not invent a deadline or claim that the prior weaknesses have been cured. A non-reliance notice, material revision or evidence that current-period analysis is unreliable would require a new eligibility and financial review.

Year-end results, expected in early 2027 with an exact date not established in the saved sources, will test the implied Q4 revenue and adjusted-profit ranges. The stated expectation of about $9B of customer prepayments and deposits over the next three months creates a separate cash checkpoint around February 10, 2027. The amount received, restrictions and linked delivery obligations matter together. A gross cash receipt without its associated terms does not settle the shareholder-value question.

A lower quote near $47 would create a roughly 30% discount to the present base value. We would still need credible capacity delivery, manageable financing and evidence that cash conversion is improving. Conversely, missing the raised operating plan, net interest exceeding the guided range, material customer-funding shortfalls or worsening reporting reliability would reduce confidence even at a lower price. The next packet should grade these exact signposts and revise the model, rather than preserve the current target after the underlying facts have changed.

Sources, history and reproducible definitions

All operating amounts are U.S. dollars in millions unless a per-share, percentage, power or share-count unit is identified. The nine-quarter series uses the latest available matching-quarter earnings-release comparisons. Earlier releases were presented in thousands and later ones in millions; the saved dataset records the scaling and original rows. Rounding can cause small differences across documents. Full income statements, reconciliation tables and available balance-sheet histories are retained rather than replacing missing fields with estimates.

Three annual cash-flow histories come from the audited 2025 Form 10-K and SEC company facts tied to that filing. CFO less cash capex is computed consistently from cash-flow statement lines, not confused with total capital additions. Preferred dividends were $58M in 2024 and $29M in 2025; they are not common dividends. Separate annual common-buyback amounts and earlier employee counts were not reliably identified in the saved summary dataset and remain unavailable rather than silently set to zero. The 2025 report gives 2,189 employees.

The fixed share denominator makes valuation routes comparable, but does not remove dilution. For any fixed modeled equity value, raising shares from 559M to 600M reduces per-share value by 6.8%; 650M reduces it by 14.0%, before considering proceeds or changes to debt. Those are mechanical sensitivity examples, not forecasts of issuance. Converting debt into shares would require reducing the corresponding debt claim as well; adding dilution while retaining converted debt would count the same financing claim twice.

The prior local Q2 episode metadata was available, but no prior research packet signpost document was found in the inspected canonical episode directory. We therefore do not manufacture a prior scorecard or claim forecast accuracy. The current dated signposts establish the next auditable comparison. Market snapshots and target-news reports are provider data, not issuer filings; stale or incompatible estimate fields are expressly excluded from the valuation. Source bytes, dates, hashes and calculation code are preserved for independent review.

• submissions_archive.json
• 10-Q_2026-03-31.html
• 10-Q_2026-06-30.html
• 10-K_2025-12-31.html
• 10-Q_2025-03-31.html
• 10-Q_2025-06-30.html
• companyfacts.json
• 10-Q_2025-09-30.html
• release_2026-02-26.html
• release_2026-08-11.html
• release_2026-05-07.html
• release_2025-11-10.html
• release_2025-08-12.html
• release_2025-05-14.html
• coreweave3q26earningspress.htm
• submissions_final_checkpoint.json
• Market: FMP quote, daily close history, ratings, target-news and annual estimate snapshots retrieved November 11, 2026. Incompatible consensus metrics are archived but excluded.