1 Artificial Intelligence (AI) Stock That Could Surprise Investors in 2026

Revenue surged from $229M in 2023 to $5.1B in 2025 with management forecasting at least a 235% jump in 2026, and a contracted backlog of $66.8B. Trailing-12-month net loss reached $1.2B, CoreWeave spent $8.2B on capex in Q4 and plans at least $30B in 2026, raising concerns about profitability and asset utilization if AI demand cools. Shares are down 55% from their 52-week high and Microsoft accounted for 67% of 2025 revenue, highlighting customer concentration risk.

Analysis

The market is treating CoreWeave as a pure capex call rather than a layered services business; that conflation understates two optionalities: (1) the ability to reprice scarce GPU/slot access in a tight supply environment, and (2) capture of higher-margin orchestration and software revenue as customers move from spot rentals to managed, SLA-backed deployments. If utilization remains above mid-teens percentage points versus industrial peers, each incremental percentage point drives disproportionate gross-margin improvement because incremental slot-revenue carries minimal additional fixed-cost increases.

Second-order winners include power utilities, regional equipment vendors (cooling, transformers) and thin-stack GPU brokers who will monetize arbitrage between hyperscaler inventory and on-demand customers; losers are large hyperscalers’ internal build programs and generic colo REITs that compete primarily on price rather than GPU specialization. Watch GPU vendor cadence and warranty/maintenance supply chains — any tightening accelerates pricing power for specialized cloud providers and compresses vendor bargaining leverage.

Key tail-risks are demand-rotation (AI spend reallocation away from training toward edge/ops), rapid vertical integration by hyperscalers, and capital-market funding stress that makes incremental capacity financing punitive. These risks play out over 6–24 months; a meaningful reversal would show up first as falling utilization and widening cash-conversion gaps rather than headline revenue misses. Near-term catalysts that will re-rate the stock are sustained utilization beats, multi-year contracted take-or-pay conversions, or a financing that preserves growth without severe dilution.

Contrarian read: current pricing discounts structural scarcity and software upsell; the market is over-indexed to headline capex quantum and underweight convertibility of backlog into sticky, higher-margin revenue. That asymmetry favors option-like exposures where upside is convex if the AI hardware cycle stays tight and utilization normalizes within 12–18 months.

Key Decisions for Investors

• Core position (tactical, 1–2% NAV): Buy CRWV equity with a 12–18 month horizon and size modestly. Hedge tail downside by purchasing 12-month 30% OTM puts to cap downside to ~30%; objective return 100–200% if utilization and convertibility catalysts hit within 12 months.
• Convexity play (speculative): Buy CRWV 18–24 month LEAP calls and finance by selling 3-month covered calls rolled monthly (buyJan-2028 calls / sell near-term calls). This preserves upside optionality while monetizing near-term implied vol; target 3:1 upside/downside skew if backlog converts.
• Relative-value observe-and-react (hedged): Pair trade long CRWV / short a broad AI hardware exposure (e.g., NVDA calls sold or reduced NVDA long exposure) sized 0.5% NAV to isolate neocloud pricing vs raw GPU price risk. If GPU supply tightens, close the short; if AI demand shocks, unwind the pair symmetrically within 90 days.
• Risk management: Set utilization and gross-margin weekly alerts; reduce CRWV exposure by 50% if utilization falls >10 percentage points quarter-on-quarter or if signed-contracted revenue conversion rate misses guidance by >20% over two consecutive quarters.