Goldman Sachs just ran some ugly numbers on the SaaSPocalypse—and found hedge funds are dumping software and piling into semis

Goldman data show a major rotation out of software and into semis: software is down 14% YTD, hedge funds have cut software to their lowest long weight since 2019, and mutual funds are at their widest underweight ex-Microsoft since 2012. Hedge funds added LRCX, AMAT, and ASML on net in Q2, while both hedge funds and mutual funds trimmed Microsoft last quarter. The article argues that AI value is increasingly accruing in semiconductors and infrastructure rather than the SaaS application layer, creating a sector-level headwind for software multiples.

Analysis

The positioning signal matters more than the relative performance gap: hedge funds are still running elevated gross/net exposure while actively reallocating away from software, which implies this is a conviction rotation rather than a simple risk-off trade. That makes the near-term path of least resistance worse for software because multiple compression can persist even if fundamentals stabilize, especially when buy-side benchmarks start treating AI exposure as a semis/infra problem instead of an app-layer problem.

Second-order winners are the picks-and-shovels names with direct AI capex linkage and visible backlog conversion, but the move likely broadens beyond pure semis into the adjacent tooling, test, and fabrication ecosystem as investors seek cleaner monetization. The more interesting loser is not just MSFT: it is any software platform whose data moat is weak enough that customers can demand outcome-based pricing and shorter contract duration. That pressure should show up first in renewal terms, then in NRR, and only later in headline revenue growth.

The contrarian setup is that the market may be overpricing the speed of value capture by semis and underpricing the lagged monetization of workflow owners. If agentic AI raises labor productivity inside enterprise software before it fully commoditizes the interface layer, the incumbents with proprietary data and embedded workflows can re-accelerate margins without needing old-school seat growth. The risk is time horizon: semis can continue to work for quarters, but software could become a better relative trade once capex growth normalizes or hyperscaler spending slows on any margin scrutiny.

Goldman’s own internal tension is the key catalyst watch. If AI productivity starts to show up in customer support and dev tools first, that is a delayed positive for software adoption but a near-term negative for pricing power. The next inflection likely comes from Q2/Q3 renewal commentary and guidance, not from broad macro data, so the trade should be managed around earnings and large-vendor contract disclosure rather than waiting for a clean sector-wide drawdown.

Key Decisions for Investors

• Stay long LRCX/AMAT/ASML vs short NOW/MSFT for 4-8 weeks: favors direct capex beneficiaries over software names exposed to renewal pressure; target 10-15% relative outperformance, stop if software guidance inflects upward on AI attach rates.
• Buy MSFT put spreads into the next earnings window: the market is still paying for AI defensiveness, but that premium is vulnerable if Copilot monetization remains gradual; prefer 3-6 month maturities to capture multiple compression rather than a one-day event.
• Pair trade long TRI vs short a basket of high-multiple SaaS: proprietary data and workflow depth should outperform interface-only software as customers demand measurable outcomes; expect 5-8% spread if contract scrutiny rises through the next two quarters.
• Fade INTC/SITM strength on rallies unless order visibility improves: this is the more crowded part of the rotation and has less direct AI moat than ASML/LRCX; better entry is after a 1-2 week consolidation, with tight risk control.
• If semis continue to rip into the next 2 earnings cycles, rotate half the position into software with actual data moats and embedded distribution rather than broad SaaS beta; that is the highest convexity contrarian exposure if the market starts to punish capex enthusiasm.