2 Stocks to Buy That Stanley Druckenmiller Added While Exiting 5 Chip and Photonics Stocks

Stanley Druckenmiller’s Duquesne Family Office boosted its Amazon stake from 45,800 shares to 541,600 shares (+1,000%+) to about $129M at June 30, and added call options on another 459,300 shares. It also initiated a new Alphabet position of 336,300 shares (~$120M), while fully exiting Micron, Intel, Broadcom, Lattice Semiconductor, and Coherent and instead adding/adding to AMD (72,900 shares), Lam Research (43,600 shares), and topping up TSMC. The filing suggests a tilt toward “AI build-out” beneficiaries tied to the biggest capital spenders—Amazon and Alphabet—supporting a mildly bullish read-through for AI-linked equities.

Analysis

The useful read-through is not “sell semis, buy tech,” but a rotation from exposed vendors to the firms that control AI demand and can internalize more of the stack. AMZN and GOOG can amortize massive capex across cloud, ads, and software; that makes their spending more durable than the order book of merchant suppliers, especially where product refresh cycles and pricing power are weak. The second-order winner is the bottleneck layer: TSM and LRCX benefit if hyperscaler spend stays elevated, while memory, legacy CPU, photonics, and smaller AI-adjacent names face a tougher mix of slower unit growth and margin pressure.

The market should be careful not to overread a lagged 13F as a clean sector call. The exits in MU/INTC/LSCC/COHR are consistent with pruning names that have the most earnings sensitivity to a capex rollover, but they do not invalidate semis broadly; in fact the new AMD, TSM, and LRCX positions suggest preference for names with either share gain or infrastructure leverage. Over the next 1-3 months, the key catalyst is whether AWS and Google continue to raise capex while maintaining cloud margins; if margins compress faster than revenue expands, the “buy the builders” thesis loses force.

Contrarian view: the consensus may be too eager to extrapolate hyperscaler capex as permanent demand. If AI monetization lags, AMZN and GOOG could face multiple compression because the market will eventually question ROI on $400B+ of combined spend. Falsifiers are simple: capex guides down, cloud growth decelerates, or AI supply constraints ease enough to reprice the whole chain lower in 2025-26.

Key Decisions for Investors

• Long AMZN vs short MU or INTC for 1-3 months: express the view that demand owners outperform price-takers if AI capex stays sticky; invalidated by a cloud-margin miss or capex cut.
• Long GOOG vs short SOXX/SMH on a 3-6 month horizon: prefer the hyperscaler that can self-fund infrastructure and monetize search/ads over a broad basket of AI suppliers; cover if semis regain leadership on a strong guide-up cycle.
• Add TSM and LRCX on pullbacks as the cleaner supply-chain winners, but size modestly: they benefit only if hyperscaler capex remains elevated and export/regulatory risk stays contained.
• Avoid chasing AVGO/LSCC/COHR on this filing alone; use them as watchlist names for a bounce only if next earnings show design-win conversion and improving backlog conversion, otherwise the risk/reward is worse than AMZN/GOOG.