Is Wall Street in a 'boom loop' or heading for a doom loop?

Wall Street is being driven by very high earnings expectations, with Q1 U.S. EPS growth now projected at nearly 28% versus 14.4% on April 1. The S&P 500 dividend yield is just 1.1%, the equity risk premium is negative and near the 2024 low of -0.7%, and AI capex forecasts were raised to more than $800 billion this year and $1.1 trillion next year. The article argues these conditions could support a continued AI-led bull market, but also raise valuation and funding concerns.

Analysis

The market is increasingly pricing a self-reinforcing loop where capex begets earnings, earnings justify higher multiples, and higher multiples lower the cost of capital for the same spenders. The non-obvious risk is that this loop is very concentrated: a narrow set of hyperscalers is underwriting both the index-level earnings revision cycle and a large share of capex demand, so any slowdown would hit earnings expectations, supplier orders, and sentiment simultaneously. That makes the current setup more fragile than a simple “expensive market” narrative suggests.

The clearest second-order effect is that AI capex is acting like an industrial stimulus program for a specific set of equipment, power, networking, and construction beneficiaries, but with delayed payoff and uncertain monetization. If ROI disappoints, the market will not just re-rate tech; it will also reprice the adjacent ecosystem that has already capitalized on the spend assumption. In that scenario, the pain likely shows up first in suppliers with cyclically inflated backlogs and second in the broader market via negative earnings revisions.

The contrarian read is that the biggest immediate risk is not valuation compression from rising yields, but narrative saturation: when buy-side consensus becomes anchored to ever-higher AI capex estimates, incremental good news loses marginal impact while any miss creates a large air pocket. The fact that dividend yields are near historic lows and ERP is negative means equities are already discounting a very favorable terminal outcome; the market has limited cushion if 2025-2026 spend stops accelerating or if monetization evidence remains intangible. Time horizon matters: this is less a one-week macro trade than a 3-12 month positioning hazard.

For MS and GS, the setup is mixed: both can benefit from underwriting, M&A, and capital markets activity if the boom persists, but they are also exposed to a late-cycle reversal in sentiment that would hit trading and financing appetite. They are better viewed as high-beta proxies on the durability of the narrative than as pure beneficiaries of AI spend itself. A sustained deterioration in the capex thesis would likely hit them via broader risk appetite before it shows up in direct fundamentals.

Key Decisions for Investors

• Reduce gross exposure to megacap AI capex beneficiaries on strength; if the theme persists, keep only core longs and sell upside via call overwrites on a 3-6 month horizon to monetize elevated implied volatility.
• Pair trade: long utility/electrical grid beneficiaries vs short AI capex suppliers that have already re-rated on orders alone; use a 6-12 month horizon and focus on names with the most stretched EV/sales versus backlog visibility.
• For MS and GS, stay tactically long only on evidence of sustained ECM/M&A and underwriting momentum; otherwise hedge with index puts or a short QQQ overlay because their upside is tied to sentiment persistence, not just fundamentals.
• If AI capex estimates keep rising without matching revenue monetization, initiate bearish exposure in the most crowded AI infrastructure names via put spreads 3-6 months out; risk/reward improves if the market starts penalizing ROIC uncertainty.
• Watch for a reversal trigger: any guidance cut from hyperscalers or a sharp deceleration in capex growth; that would justify de-risking cyclical tech and reducing beta immediately, as the unwind could be fast and correlation-driven.