Down 40%, CoreWeave Is Being Left Behind By the Market

CoreWeave (CRWV) is down 40.57% over the past year despite Q1 2026 revenue of $2.08B (+111.69% YoY) and a $99.4B backlog, as net loss widened to $740M and capex surged to $7.7B in one quarter. Gross margin fell from 78% to 68% and adjusted operating margin slipped to ~1%, with free cash flow at -$4.7B and interest expense up to $536M. The stock also faces a securities fraud class action related to alleged data center construction delays; analysts nonetheless see a $142.29 average price target (+53.83% vs. current), contingent on margin recovery to low double-digit levels by Q4 2026 and execution against a $50.8B liabilities/debt burden.

Analysis

The market is effectively separating “capacity owners” from “cash-flow compounding” in AI infrastructure. CRWV is being priced like a leveraged utility whose growth must be continuously refinanced, while NBIS/IREN are earning a scarcity premium because they are proving they can translate demand into operating leverage faster. In the near term, that favors names with visible EBITDA conversion and lighter incremental capex; over the next 1-3 months, the main catalyst is not backlog chatter but whether credit spreads or refinancing terms improve enough to reduce dilution risk.

The second-order issue for CRWV is that backlog is only valuable if it clears the capex and interest hurdle before the balance sheet forces a reset. A business that requires massive upfront buildout and still shows margin compression becomes highly sensitive to any delay in customer ramp, power delivery, or deployment timing; that creates asymmetric downside if even one large customer slips. By contrast, NVDA remains the cleaner beneficiary of the AI buildout because it captures the capex cycle without taking construction, financing, or utilization risk.

Contrarian view: the selloff may be partially overdone if management can demonstrate that the current capex wave is a one-time land grab rather than a permanent burn rate, and if lenders continue to fund against contracted revenue. But the burden of proof is now on execution metrics, not pipeline size. What would falsify the bearish view is a clear turn in unit economics: lower capex intensity, sustained gross margin stabilization, and evidence that operating margin can inflect without new equity capital.

Key Decisions for Investors

• Short CRWV vs long NBIS as a 1-3 month relative-value pair: NBIS has already demonstrated EBITDA conversion, while CRWV still trades like a balance-sheet story; target is further multiple dispersion if financing concerns persist.
• Long NVDA vs short CRWV for the next earnings cycle: same AI demand theme, but NVDA monetizes the buildout with far less leverage and capex risk; best if AI spending stays firm but credit tightens.
• Avoid chasing CRWV strength into any backlog headline; wait for proof in gross margin and FCF before adding long exposure. If those metrics do not improve by the next quarter, the equity remains vulnerable to dilution-driven repricing.
• If already long CRWV, hedge with downside puts into the next financing or litigation milestone; the cleanest failure mode is not demand collapse, but a funding event that resets the equity multiple.
• Watch the AI infra credit market: if spreads widen or a refinancing window closes, reduce CRWV exposure immediately; that is the highest-conviction catalyst that can reverse the current underperformance.