Billionaire Stanley Druckenmiller Sells Micron and Is Piling Into This Other Unstoppable Artificial Intelligence (AI) Chip Stock Instead

Duquesne Family Office (per its Q2 13F) fully exited Micron (MU) after a +231% 2026 rally to a ~$1T market cap, reallocating to Advanced Micro Devices (AMD). The article cites AMD Q2 revenue of $11.5B (+50% YoY) with data center revenue of $6.7B (> doubled) near 60% of sales, alongside guidance for continued acceleration in 2H. While AMD shares are up >120% and trade at a ~63x forward P/E versus ~26x for semis broadly, the piece frames the rotation as risk-managed participation in AI compute demand rather than a pure-play memory bet.

Analysis

This rotation reads less like a broad AI endorsement and more like a preference for scarce, durable compute over memory that can be bid up faster than fundamentals. The second-order implication is that the market may start rewarding names with clearer share-gain runways and penalizing AI adjacencies where supply can normalize quickly. If that re-rating continues, the real relative losers are not just MU but also other memory-linked suppliers that have benefited from a late-cycle scarcity narrative.

AMD is the cleaner beneficiary if hyperscalers keep diversifying away from a single supplier, but the stock already discounts a meaningful amount of execution. The setup is fragile: any slip in accelerator ramp, gross margin, or data-center mix will compress the multiple quickly because the current valuation leaves little room for a one-quarter miss. By contrast, the bullish case extends over 6-18 months if AMD keeps taking incremental wallet share; that is where the operating leverage matters most.

The contrarian view is that the move may be overdone in both directions. MU can still be a cash-flow machine if HBM tightness persists longer than expected, while AMD may be the kind of crowded “next-best” AI trade that gets sold on good news once the filing-induced chase fades. The key near-term catalyst is the next earnings print and hyperscaler capex commentary; if AMD guide-ups again, momentum can extend, but if memory pricing holds and AMD’s growth decelerates, the relative trade unwinds fast.

Key Decisions for Investors

• Prefer a relative-value long AMD / short MU pair for 1-3 months: thesis is compute scarcity vs memory mean reversion; stop if MU HBM pricing and guide remain tight or if AMD data-center growth decelerates.
• Avoid initiating fresh outright longs in AMD at current valuation without confirmation; use post-print pullbacks or a failed breakout above recent highs as the better entry point, since the multiple can compress 15-20% on a small execution miss.
• If you want AI beta with less single-name risk, own SMH or SOXX instead of chasing AMD alone; this captures the capex trend while reducing the penalty from any one product ramp hiccup.
• Set an alert on MU for the next two quarters: if HBM and DRAM pricing stay firm, the 'sell memory, buy compute' trade is likely crowded and should be covered.