This could infrastructure stock has been hit hard of late. UBS says to buy it

UBS initiated CoreWeave with a buy rating and $120 price target, implying 38% upside from Tuesday's close, arguing that leverage and credit-risk concerns are likely peaking. CoreWeave shares have fallen nearly 18% over the past three months, with the decline worsening after its planned $3 billion convertible-debt offering and new-share issuance raised dilution concerns. UBS expects sustained enterprise AI-compute demand and GPU pricing to lift revenue per GW from about $11 billion currently to more than $15 billion over time.

Analysis

The key underwriting issue is not demand but whether incremental GPU capacity earns a return above an increasingly expensive cost of capital. Higher revenue density per GW would improve fixed-cost absorption, but it does not automatically translate into equity value if depreciation, power costs, financing expense, and recurring hardware-refresh capex rise in parallel. CRWV should therefore trade more like a leveraged infrastructure/credit hybrid than a clean AI software beneficiary until free-cash-flow conversion becomes visible.

The convertible issuance creates a near-term technical overhang: arbitrage desks may hedge purchased converts through stock sales, while new equity expands the effective float. That can suppress the shares for weeks after pricing even if the financing meaningfully extends liquidity runway. The more important signal is the conversion premium, coupon, maturity, covenant package, and whether lenders require collateral or tighter restrictions; favorable terms would validate that credit risk is easing, while punitive terms would make a nominally growth-accretive buildout equity-destructive.

Consensus sponsorship limits the informational edge in a simple long. The contrarian opportunity is that the market may be pricing a binary refinancing outcome while contracted enterprise demand and operational execution reduce that binary risk; conversely, bulls may be underpricing customer concentration and GPU obsolescence risk. Over the next 1-3 months, financing terms and any backlog-to-revenue conversion evidence matter more than another bullish analyst target; over 6-18 months, the decisive metric is whether EBITDA growth outpaces net debt and maintenance/refresh capex.

A sustained rerating requires management to demonstrate improving leverage on a net-debt-to-EBITDA basis and stable or rising gross margin despite new capacity coming online. Thesis failure would be a guidance reduction, materially weaker revenue-per-GW realization, rising funding costs, or evidence that hyperscalers and other GPU-cloud providers are using excess capacity to compress rental rates.

Key Decisions for Investors

• Treat CRWV as a post-convert-pricing event-driven long, not an immediate full-size position. Initiate only after final terms show a conversion premium consistent with limited dilution and no unexpectedly restrictive secured debt features; target a 3-6 month move toward the $120 analyst valuation, with position risk cut if financing terms imply materially higher cash interest or incremental equity issuance.
• Use defined-risk upside rather than uncovered equity exposure while the convertible hedge flow clears: buy 3-6 month CRWV call spreads with a long strike near spot and short strike near $120, sized for a maximum premium loss. This captures a de-risking rerating while limiting the downside from a credit or guidance shock.
• Monitor CRWV debt/convert secondary pricing and credit spreads daily for the next 4-8 weeks. Tightening spreads alongside stable equity would support adding exposure; widening spreads despite a stock bounce is a warning that the equity move is technical rather than fundamental.
• Do not pair CRWV against a broad AI semiconductor short. A better hedge for a long is modest downside protection through CRWV puts or reduced gross exposure, because the principal risk is issuer-specific capital structure and execution rather than a generalized AI-demand downturn.