Goldman's Kaplan on Fed's Warsh and AI Capex Boom

Robert Kaplan expects Kevin Warsh to take a more neutral tone in his first news conference as Fed chair, a modestly supportive signal for policy continuity. Kaplan also described an "historic" U.S. capital expenditures boom, which points to strong business investment momentum. The remarks are mostly commentary rather than new policy action, so immediate market impact should be limited.

Analysis

A more neutral Fed tone lowers the left-tail risk of a policy mistake that could choke off the capex cycle before it becomes self-funding. The first-order winner is less about rate-sensitive beta and more about firms with long-duration investment pipelines: semis, industrial automation, datacenter infrastructure, and grid/electrification suppliers should see project hurdle rates stabilize, which tends to unlock backlog conversion with a 2-4 quarter lag.

The bigger second-order effect is on capital allocation. If management teams believe the Fed will tolerate growth and inflation runs a bit hotter, buybacks become relatively less attractive versus capacity expansion, especially for companies sitting on large cash balances or exposed to AI/data center demand. That shifts relative value toward equipment, tools, and electricals over capital-return-heavy financials and mature defensives.

For GS specifically, the setup is modestly positive but not explosive: a neutral Fed tone supports ECM, M&A psychology, and trading activity, but the real upside would come only if lower volatility and better risk appetite drive a broader underwriting rebound. The more interesting read-through is that an “historic” capex boom is a late-cycle signal in disguise: it can extend the cycle for months, but it also raises the odds of bottlenecks, margin pressure in upstream suppliers, and a later disinflation impulse if productivity gains outrun wage growth.

Contrarian view: consensus may be underestimating how concentrated the capex boom is. If spending stays AI/cloud/electrification-heavy rather than broad-based, the market may overpay for cyclical industrial beneficiaries while missing that the true winners are the infrastructure bottlenecks — power, cooling, networking, and memory/compute supply chains. A neutral Fed helps, but the trade still depends on execution; if capex orders fail to convert into revenue within two quarters, enthusiasm can fade quickly.

Key Decisions for Investors

• Long XLK / short XLI for the next 3-6 months: own the spend beneficiaries with pricing power and short the broader industrial basket that is more exposed to second-order margin pressure if capex remains concentrated rather than broad-based.
• Add to semicap equipment and grid-electrification exposure on pullbacks over the next 1-2 weeks: AMAT/LRCX and ETN benefit from a lower policy-risk backdrop and can re-rate if order visibility improves through the next earnings season.
• Initiate a tactical long GS into the next 30-60 days via call spreads: neutral Fed tone should support advisory and underwriting sentiment, but keep size modest because the upside is more cyclical than structural.
• Pair long infrastructure bottlenecks / short mature capital-return names over 3-6 months: favor firms tied to power, cooling, and data-center buildout over cash-return stories that may see allocation pressure if managements pivot from buybacks to capex.
• If semis rally into the narrative before bookings confirm, fade with defined-risk puts: the trade can reverse quickly if capex turns out to be front-loaded or overly concentrated in a handful of hyperscalers.