tkalAI Capital Flows
Earnings deep dive · Synthesis edition · AI Cloud & Infrastructure

CRWV Q2 2026 — The Customer Was Never the Risk. The Lender Is.

Here is the whole print in one sentence: revenue grew 112% and interest expense grew 140%. Everything else is commentary. CoreWeave doubled its top line to $2.58B, compounded its backlog past $104B, and I am asking you to stop looking at all of it — because the demand question was settled quarters ago and it is not what prices this equity. What prices this equity sits below the EBITDA line, where $1.393B of depreciation and $640M of interest turned a $1.51B profit into a $626M loss. This is a demand-long, balance-sheet-short company, and the one number that decides its next year is the one management controls least: the spread on the next financing.
Published · Post-print synthesis (reported after close ) · CRWV · Committed analyst read — not advice · For analysts covering AI infrastructure, neoclouds, hyperscaler capex, and the data-center credit stack

The only question that matters

CoreWeave is the purest thing the public market has built to express the AI-infrastructure trade, and that purity cuts both ways. Shares closed at $90.32 on into the after-close release, roughly flat on the month with the options market pricing a wide move and 30-day IV rank near the 51st percentile (unverified — options data, not filing figures). Read the setup honestly and the bull case needed no defending: the backlog was already historic, the customer list already read like the frontier-AI directory. So the bears did the sensible thing and narrowed their whole argument to one place — below adjusted EBITDA. That is the correct place to look, and this piece is going to stay there, because the carrying cost is senior to the shareholder: depreciation and interest get paid in full before the equity sees a dollar. Days before the print, a fresh $2.6B loan put the debt back in the headlines and JPMorgan nudged its target to $110 from $105 (unverified — analyst data). Notice what moved the stock in the run-in. Not a customer. A loan.

Read the two numbers side by side

Management leads with growth, so let us take the growth at face value and then read what sits underneath it. Revenue landed at $2.575B, up 112% year-over-year from $1.212B — an annualized run-rate near $10.4B, and a beat on the ~$2.56B consensus. Genuinely excellent, and genuinely not the point. Now put the second number next to it: net interest expense of $640M, up 140% from $267M. Decode that pairing, because it is the entire thesis in two figures — the cost of the balance sheet grew faster than the business did. A company can double revenue every year and still walk backwards if the line that funds the doubling compounds quicker, and this quarter it did. That is why the headline growth rate is not the growth rate the equity is priced on.

The rest of the P&L simply narrates that collision. On the leverage side: adjusted EBITDA of $1.510B at a 59% margin (down from 62%), and adjusted operating income of $128M at a 5% margin — collapsed from 16% a year ago as the fleet scaled and depreciation stepped up. On the GAAP side, once you pass $1.393B of depreciation & amortization and that $640M interest bill, the $49M operating loss becomes a net loss of $626M, or $(1.14) per share (versus $(0.60) a year ago; ~$(1.20) expected). Adjusted net loss was $(567)M. Do not skip past the margin line — 16% to 5% adjusted operating margin in a year is management telling you, in its own preferred non-GAAP language, that the fleet is getting more expensive to carry faster than it is getting more profitable to run. The bridge from $1.51B of EBITDA to a $626M loss is not an accounting quirk. It is the company.

Below adjusted EBITDA: where the $1.51B profit goes CoreWeave Q2 2026 · adjusted EBITDA vs. the costs beneath it · US$, bars to scale Adjusted EBITDA +$1.510B 59% margin BELOW THE LINE — THE DRAGS Depreciation & amort. $1.393B Interest expense $640M up 140%, from $267M WHERE IT LANDS GAAP net loss $(626)M $(1.14)/sh Adjusted net loss $(567)M
The two drags below adjusted EBITDA — $2.03B combined — outweigh the $1.51B of EBITDA, which is why a 59%-margin quarter still prints a $626M loss. Adjusted EBITDA excludes D&A and interest; the GAAP net loss also reflects other items and tax, so this shows the major drags, not a full line-by-line reconciliation.
Paid subscribers

The rest of this dive is for paid subscribers.

The headline numbers are above. The 9 sections below carry the mechanism, the peer read, the valuation work, and the dated tests that decide it.

  • The demand is real — and it is not the argument
  • The guide answers the demand question and enlarges the real one
  • Follow the money that is actually funding this
  • How the rest of the complex should read this
  • The valuation depends entirely on which number you believe
  • The strongest case against me
  • What decides it from here
  • The trade (analysis, not advice)
  • Bottom line
Unlock the full dive →

Already a subscriber? Sign in.

Subscribe · free weekly

This is one print. The pipeline runs every week.

Every Monday, the single sharpest call from the pipeline — a dated, falsifiable read on credit, demand, and capex, then graded in public on the scorecard. Free.

One email to confirm. Unsubscribe any time.

Sources