Coreweave’s McBee sells $13.5m in shares

CoreWeave CDO Brannin McBee sold 166,350 Class A shares on April 6 for $13.54M at $80.3825–$81.8822 (stock now $88.86, ~82% 1yr gain) and converted 150,000 Class B shares into Class A. The company reported strong MLPerf Inference v6.0 results for GB200/GB300 systems and closed an oversubscribed $8.5B delayed-draw term loan facility. Analyst views are mixed—Evercore ISI and Citizens reiterated outperform ratings ($120 and $180 targets), Barclays stayed Equalweight at $90, while Bernstein reiterated Underperform at $56—providing divergent signals on valuation and competitive outlook.

Analysis

Specialized GPU-cloud providers and their hardware partners sit at the intersection of two forces: escalating GPU unit demand and fierce pricing pressure from hyperscalers. On a multi-quarter view this amplifies winners who can vertically integrate firmware, systems and procurement (favoring boutique OEMs) while penalizing balance-sheet-constrained operators that must monetize growth through margin-sacrificing contracts.

Recent financing and equity-liquidity dynamics materially change optionality: robust lending appetite reduces near-term refinancing risk but raises the probability of covenant-driven project prioritization and governance compromises over 6–18 months. Benchmarks and product-cycle milestones remain the most actionable catalysts in the coming weeks; contract disclosures and partner shifts will move sentiment rapidly and can re-rate multiples by 20–40% within a quarter.

The durable tail risks are loss of preferential GPU allocations to hyperscalers, rapid platform obsolescence when next-gen accelerators arrive, and concentrated customer cancellations; any one could flip a growth multiple to a multiple more typical of hardware capex cycles within 12–24 months. Contrarian upside exists if smaller clouds sustain premium pricing for specialized workloads — that outcome requires sustained non-linear demand from a narrow set of enterprise AI applications and predictable GPU supply beyond current allocations.

Key Decisions for Investors

• Pair trade (6–12 months): Long SMCI equity (or buy 12–18 month 10% OTM call spread) while short CoreWeave via a 12–18 month put spread financed by selling a smaller amount of covered calls — structural bull on systems integrators vs execution/valuation risk at specialized cloud operators. Target asymmetric payoff ~2:1 upside vs max loss capped by spreads.
• Tactical options (0–3 months): Buy NVDA near-term earnings-protective call diagonal (buy 3–6 month calls, sell 1–2 month calls) to capture continued GPU pricing tailwinds while harvesting premium. Risk: gamma into earnings; reward: leverage to multi-quarter capacity tightness.
• Event-driven short (3–9 months): Buy CRWV 3–9 month put spread to express valuation and execution downside ahead of major contract renewals or partner updates. Size as 1/2 of a single-name long to keep fund-level convexity balanced; expected payoff 3–5x if adverse contract news hits.
• Credit/alpha trade (6–18 months): Favor banks with underwriting access to syndicated tech facilities (selective long BCS or similar) on a tactical basis to capture fee and spread cycles as tech capex financings reprice; hedge duration and macro beta. Reward: capture excess spread and fees; risk: sudden widening if tech capex falls sharply.