Billionaire David Tepper of Appaloosa Tripled His Stake in Micron but Has Been Dumping a Historically Cheap AI Stock Over the Last Year

Appaloosa manager David Tepper purchased 1,000,000 Micron shares in Q4, raising the fund's Micron stake to 1.5M shares and making it the fund's 4th-largest holding; Micron’s fiscal Q1 gross margin rose to 56% from 38.4% a year earlier and the stock traded at a forward P/E of ~9 (as of Mar 13). Simultaneously, Tepper sold 1,312,069 Alibaba shares in Q4, trimming his Alibaba position by 57% in 2025, citing profit-taking and possible concern over US-China trade policy; Alibaba’s forward P/E is ~15. These filings are likely to move the individual stocks moderately rather than broader markets.

Analysis

Micron’s current tailwind is less a pure demand story than a temporary market structure advantage: concentrated GPU/HBM supply + multi-year cloud inventory restocking has created a pricing regime that can sustain materially above-cycle margins for 6–18 months. That window simultaneously incentivizes aggressive capex from competitors and third-party subcontractors, so the durable outcome is binary — either rational, phased capacity additions keep pricing elevated or a front-loaded capex response drives oversupply and margin erosion within 12–24 months.

Second-order winners include contract OSATs and substrate suppliers that service HBM stacks; they will see margin flow-through without the balance-sheet volatility of memory fabs, creating attractive mid-cycle cash generation ahead of semiconductor equipment names that follow. Conversely, large hyperscalers are the natural marginal buyer and also the latent swing factor — if they slow procurement to manage PoP/GPU utilization, pricing power will compress quickly and GPU vendors (and their adjunct suppliers) will feel the hit within a single product cycle.

On the China/AI name side, headline discounts have likely priced in near-term geopolitical and regulatory scenarios, but they underweight optionality from fast-follow monetization of on-prem/cloud AI services and a large cash cushion. The key catalysts to watch are supplier capex cadence, hyperscaler inventory disclosures, memory ASP indices, and China stimulus signals — each can flip the thesis inside 3–12 months. Trade tactics should therefore balance asymmetric upside exposure to Micron’s pricing power with hedges against a 20–40% downside scenario if oversupply or demand softness emerges.

Key Decisions for Investors

• Initiate a tactical long in MU (size 2–4% NAV) via a calendar/LEAP spread: buy Jan-2028 LEAP calls (targeting ~12–18 month capture of structural rerate) and sell 3–6 month calls to finance carry. Target return: 40–80%+ if pricing stays elevated; downside: ~30–40% if memory ASPs collapse. Cut to half position on a 20% drawdown and exit on a >30% margin compression print from the company.
• Pair trade: long MU / short BABA equal notional (size 1–3% NAV each leg) over 6–12 months to express AI-infrastructure vs China-adjacent AI exposure. Risk management: cap the short leg by using OTM calls as insurance (buy BABA +15% calls) and tighten the pair if divergences exceed 15% intraperiod. Reward: asymmetric if HBM scarcity persists while China macro/regulatory news disappoints.
• Options hedge for portfolio tech exposure: buy MU 6–12 month puts (insurance) sized to cover 25–50% of gross long exposure to protect against a rapid destocking/oversupply shock. Cost is insurance but preserves optionality for continued upside if the pricing regime holds.
• Monitoring & triggers (no position): set alerts for (a) public capex ramps from Samsung/SK Hynix/Chinese fabs, (b) hyperscaler inventory commentary or GPU shipment changes, (c) Chinese macro stimulus announcements — treat any two of these occurring within a 90-day window as a signal to reduce gross long exposure by 30–50%.