Dan Loeb reveals a stake in a semiconductor ETF: find out more

Third Point added a new position in the VanEck Semiconductor ETF (SMH) in Q1 2026 and increased exposure to AI-linked chip names including ASML, Lam Research, KLA Corp, Broadcom, and Hut 8. The portfolio shift signals a more aggressive bet on the AI and semiconductor trade. The article is portfolio positioning news rather than a company-specific earnings or guidance update, so likely market impact is limited.

Analysis

This is a positioning signal more than a fundamental one: a respected multi-strat is adding beta to the AI supply chain at a time when the market is still rewarding “picks and shovels” over end-demand. That matters because chip-equipment exposure tends to be the second-order beneficiary of an AI capex wave, with the real earnings acceleration often showing up 2-4 quarters after hyperscaler spending commitments. The immediate implication is not just higher multiple support for ASML and LRCX, but tighter crowding in a trade that is increasingly being owned through the same factor sleeve.

The competitive read-through is that the bottleneck remains advanced process complexity, not just AI enthusiasm. If leading-edge node migration keeps accelerating, ASML’s moat deepens while wafer-fab equipment peers with weaker share in EUV-adjacent spend likely underperform on relative growth. Broadcom and HUT are more exposed to the market’s willingness to finance AI infrastructure; their upside is more sensitive to capex durability and liquidity conditions than to near-term unit demand.

The main risk is that this becomes a consensus trade too quickly. When hedge funds pile into the same AI enablement basket, returns often compress even if fundamentals remain fine, because the market starts discounting 12-18 months of good news in weeks. A catalyst that can reverse the move would be any sign of hyperscaler capex discipline, export-control escalation, or a rotation away from semis into monetization-layer software/enterprise names.

The contrarian view is that the better expression may not be chasing the highest-quality equipment names after they are already owned, but fading the more narrative-driven parts of the basket and staying long the most structurally underappreciated beneficiary. If AI spending broadens beyond a few mega-cap buyers, second-tier suppliers and infrastructure enablers could re-rate later, while the first-wave winners become vulnerable to valuation digestion.

Key Decisions for Investors

• Stay long ASML vs. short a broad semiconductor basket for 1-3 months: best risk/reward if AI capex remains intact, because ASML should retain pricing power while generic semis re-rate less efficiently.
• Add LRCX on pullbacks only, with a 6-9 month horizon: equipment order leverage should show through later, but avoid chasing after a positioning-driven move.
• Use AVGO as a tactical long only on weakness, not strength: upside is tied to continued AI networking demand, but multiple expansion is already more consensus-sensitive than ASML/LRCX.
• Treat HUT as a high-beta call option on AI infrastructure sentiment; keep size small and consider defined-risk structures because it is most exposed to financing and risk-appetite reversals.
• If semis gap higher on follow-through flow, hedge with an index put spread on SOXX/SMH for 1-2 months: crowded positioning raises near-term drawdown risk even if the secular view remains positive.