Buy the dip in this AI 'neocloud' and don't believe bubble 'conspiracy theories,' says Freedom Capital's Meeks

Paul Meeks of Freedom Capital initiated coverage of Nvidia-backed CoreWeave (CRWV) with a buy rating and $100 price target, implying roughly 17–18% upside from the recent close and forecasting a potential retest of the $153 peak within two years. CoreWeave shares plunged about 37% this quarter after it cut full-year 2025 revenue guidance following a construction delay at one of its 41 data centers, but Meeks highlights a $56 billion multi-year revenue backlog and a 6.56% Nvidia stake (worth ~ $2.1 billion) as downside mitigants. Meeks argues CoreWeave trades at ~2.8x his 2026 sales estimate versus Nebius at 7.3x, while Street consensus price target is $131 (≈51% upside).

Analysis

Market structure: CoreWeave (CRWV) and Nvidia (NVDA) are clear beneficiaries if AI infra demand sustains; hyperscalers (big spenders) win via early capacity. High-multiple neocloud peers like NBIS (7.3x 2026 revenue) are vulnerable if supply/demand normalizes while CRWV (~2.8x per Meeks) looks relatively cheap, implying potential share reallocation. GPU supply tightness and CRWV’s reported $56bn backlog signal multi-year demand but concentrate counterparty risk with a few hyperscalers and Nvidia.

Risk assessment: Tail risks include a >=20% collapse in hyperscaler AI CAPEX, a sudden Nvidia change in commercial support or export controls, or a multi-data-center construction failure; any would cause rapid re-rating. Near‑term (days–weeks) expect high volatility on news (guidance updates, Nvidia earnings); medium-term (3–12 months) conversion of backlog and capital markets access matter; long-term (1–3 years) execution, energy contracts, and pricing power determine margins. Watch thresholds: backlog decline >15% or >90‑day payment delays should trigger defensive action.

Trade implications: Favor size‑limited, idiosyncratic long exposure to CRWV (buy-on-weakness) versus short exposure to richly valued NBIS to capture relative multiple compression; express via options to control downside. Use call spreads to target 6–18 month upside and sell short-dated implied vol if IV spikes; rebalance if CRWV trades above $150 or backlog-growth stalls.

Contrarian angles: Street underweights execution and power/infrastructure constraints and overweights headline AI-bubble fear — sell‑offs likely overshoot on psychology, not fundamentals. However NVDA concentration (6.56% stake) is both a backstop and a governance risk; historical parallels (infrastructure overbuild cycles) warn that rapid capacity scaling can create localized oversupply and margin pressure.

Key Decisions for Investors

• Establish a 2–3% long position in CRWV on a pullback to $80–$95 (≈10–25% below Meeks $100 target); target partial profit at $100 within 6–9 months and trim to halve position at $150 over 12–24 months; stop-loss 30% below entry.
• Implement a relative-value pair: go long CRWV (size 1) and short NBIS (size 0.4–0.6) to neutralize macro beta; target capture of multiple convergence (NBIS down/flat, CRWV up) over 6–12 months.
• Buy a 6–18 month CRWV call spread (long ATM or slightly OTM call, sell a higher strike ~30–50% OTM) sized to 1–2% portfolio to cap premium outlay; consider selling 1–3 month calls against the spread if IV > 60% to finance.
• Reduce high‑beta cloud/neocloud exposure by 20–50% (trim names trading >5x 2026 revenue expectations) and redeploy proceeds into NVDA (buy on dips up to 3% portfolio, given strategic stake) and energy/utility names with long‑term power contracts to hedge data‑center consumption risk.
• Monitor within 30–60 days: (a) CRWV quarterly backlog conversion rate and top‑3 hyperscaler payment cadence; (b) Nvidia guidance on GPU supply and partner program changes. If backlog falls >15% or payments slip >90 days, reduce CRWV exposure to zero immediately.