Goldman: Could Make Sense to Diversify Away From Chipmakers

Goldman Sachs' Christian Mueller-Glissmann said the recent tech selloff could justify diversifying away from semiconductor makers and toward hyperscalers. He characterized semiconductors as highly cyclical, implying a more defensive positioning within technology rather than a broad bullish call. The commentary is likely to influence sector rotation views, but it does not include new fundamental data or company-specific catalysts.

Analysis

The message is not that semis are bad businesses, but that the market is paying too much for the wrong part of the AI stack at this point in the cycle. If capex growth normalizes even modestly over the next 2-4 quarters, the earnings multiple on chipmakers can compress faster than the underlying revenue grows, because expectations are already forward-loaded and inventory leverage works both ways. Hyperscalers, by contrast, have more control over deployment pace and can defend margins by throttling spend, repricing software/services, or monetizing AI features later in the cycle.

The second-order effect is that a rotation from semis into hyperscalers is effectively a bet on durability of AI demand rather than the speed of infrastructure buildout. That should help the large-platform names that own distribution, data, and customer relationships, while pressuring the smaller set of pure-play suppliers with more direct cyclicality and less pricing power. It also shifts risk down the supply chain: equipment, packaging, and memory vendors are more exposed if procurement slows, while cloud names may see near-term capex skepticism but better medium-term FCF if investors decide the build is getting more disciplined.

The main catalyst is a sequence of softer prints on orders or guidance that forces the market to re-rate growth quality, not just growth rate. The risk to the rotation trade is a fresh wave of AI capex announcements or a stronger enterprise adoption cycle that re-accelerates semiconductor demand within 1-2 quarters. Another contrarian point: if semis have already de-risked materially, the better trade may be selective long quality chips versus the broad basket, because a lot of bad news may already be in price while hyperscalers still face antitrust and capex skepticism.

Key Decisions for Investors

• Rotate from broad semiconductor exposure into hyperscalers: long MSFT/GOOGL/META basket vs short SOXX over the next 1-3 months; thesis is lower earnings dispersion and better margin control if AI capex moderates.
• Sell semis on strength using call spreads in NVDA or SOXX into any post-selloff bounce; target 4-8 weeks, with defined risk if new capex guidance re-accelerates.
• Prefer the hyperscaler leg through cash-flow resilient names: buy MSFT or GOOGL on 5-10% pullbacks as a 6-12 month compounder trade, since they can monetize AI with less balance-sheet strain than chip suppliers.
• For a cleaner relative-value expression, go long AMZN or META vs short a basket of cyclical semiconductor equipment names for 2-3 quarters; benefit if cloud monetization improves while fab spending growth decelerates.