Want to diversify away from AI? Goldman says look at these stocks

The important signal is not a wholesale abandonment of AI; it is that crowded capital is looking for duration of cash flows with lower earnings-beta. MSCI is the cleanest expression of that trade: recurring fees, limited AI substitution risk, and multiple support if institutions rotate into “quality at a discount.” MAR can participate, but it is more exposed to late-cycle consumer softness than the market is pricing; FWONK is a better scarcity asset, but the rerating case depends on continued monetization in the U.S./China rather than broad macro beta.

Second-order effects matter here: if AI leadership pauses for more than a few weeks, there is likely mechanical de-grossing in the most crowded AI infrastructure names first, then a flow into compounders and consumer-experience stocks. That favors MSCI and, to a lesser extent, MAR over the next 1-3 months, but it also means the move can reverse quickly if AI earnings re-accelerate or capex guidance stabilizes. Over 6-18 months, the risk is that “defensive quality” gets crowded and expensive, especially if rates back up and the market stops paying for stability.

Contrarian view: the consensus may be overestimating how durable the non-AI rotation is. These are not cheap cyclicals; they are premium businesses being framed as bargains, so upside is likely to come from multiple support, not explosive earnings revision. The cleaner trade is relative value, not outright longs: own the strongest recurring-revenue compounders and hedge with AI beta, rather than betting that capital permanently exits the AI complex.