Billionaire Philippe Laffont Sold CoreWeave and Bought This Artificial Intelligence (AI) Stock Instead

Coatue sold its CoreWeave stake (once its largest holding) and nearly doubled its position in Applied Materials in Q4. CoreWeave shares have dropped ~50% since October despite backlog rising from $15.1B at end-2024 to $66.8B at end-2025 and projected revenue to more than double in 2026; it remains highly leveraged and cash-flow negative. Applied Materials is positioned to benefit from chipmaker capex (TSMC $52–56B guidance; Micron >$25B expected capex) with management forecasting >20% equipment growth this year; the stock trades near 30x forward earnings and analysts expect ~25% EPS growth in 2027.

Analysis

The structural shift is not just ‘more AI demand’ — it is a change in who captures margin from that demand. Scale incumbents with broad tool portfolios and integrated R&D (i.e., companies that sell across logic and memory flows and can embed themselves in customer roadmaps) will see disproportionate margin expansion as fabs stretch multi-year build schedules and rationalize supplier bases. Conversely, customer-facing, asset-heavy build-and-lease models face a double-duration exposure: demand is long-dated (multi-year contracts) while capacity financing and execution risk are short-dated, creating refinancing and contractor-timing vulnerabilities.

Mechanically, two second-order chains are important: (1) multi-year fab ramps increase demand for back-end specialty tools (metrology, deposition, high-volume automation) and services where incumbents can cross-sell, and (2) tight GPU allocation (channelled by a few suppliers) creates episodic revenue spikes for colo providers but also concentrates counterparty risk if a contractor or supplier misses schedule. These dynamics give a 6–24 month visibility window for tool vendors on revenue and margin expansion, but only a 3–12 month window for colo providers to demonstrate free cash flow stability.

Key risks that could reverse today’s tilt are macro-driven credit tightening and an abrupt reallocation of high-end GPUs away from cloud/colo to captive hyperscalers. Watch quarterly fab-equipment bookings, large customer capex updates, and debt covenant reviews as near-term catalysts. Structurally, this is not a binary cycle call — it’s a selection call: pick firms that convert booked demand into sustainably higher margins rather than those that simply scale assets and leverage balance sheets.

Key Decisions for Investors

• Long AMAT via a defined-cost options spread: buy 12-month AMAT ATM call and sell a 24–30% OTM call (approx. 9–12 month tenor). Position size: 0.5% notional of portfolio. R/R: target +30–40% on the underlying within 9–12 months; max loss = premium paid (~100% of premium). Stop: unwind if AMAT misses two consecutive guides on bookings or margin contraction >150bp.
• Pair trade (dollar neutral): Long AMAT cash 0.6% AUM funded by short CRWV equity 0.6% AUM. Time horizon 6–12 months. Rationale: capture differential between durable tool vendor margin expansion and exposed, levered colo execution risk. Risk controls: hard stop on the short if CRWV rises >25% from entry; trim long by 25% if AMAT underperforms sector by >8% in 30 days.
• Short CRWV via put or small outright short: size 0.25–0.75% AUM, horizon 3–9 months. Thesis: refinance/execution and GPU-allocation tail risk create asymmetric downside vs upside; target -30–50% if credit scarcities surface. Hedge: buy a cheap call to cap gap risk or pair with long AMAT to reduce directional beta.
• Event hedge: buy protection on high-yield/asset-backed datacenter credits or increase cash liquidity to 1–2% tactical for opportunistic add if contractor/financing headlines trigger widespread markdowns. Time horizon: 0–6 months; payoff: large asymmetric protection if market re-prices levered infra builds.