Bruce Yandle: America chugs along in a ‘K-shaped’ economy

The U.S. economy is exhibiting a pronounced K-shape: ADP reported a 32,000 drop in November payrolls with small firms (1–49 employees) laying off about 120,000 while larger employers added staff, even as sectors such as health care, education and information remain strong. Massive investment in data centers and associated power infrastructure is materially boosting GDP — a Harvard study by Jason Furman finds that without technology-related infrastructure 2025 H1 real GDP would have been about 0.1% instead of 2.2% — supporting strong GDPNow estimates (Atlanta Fed ~3.9% for Q3) despite weak manufacturing, construction and labor-market breadth, rising student debt burdens and disruption from AI. Investors should watch data-center capex, energy/power demand, and sectoral exposure as drivers of growth that may not translate into broad-based labor-market recovery.

Analysis

Market structure is bifurcating: hyperscalers and data-center owners (EQIX, DLR, SWCH) plus grid-facing utilities (NEE, NRG) are net beneficiaries as large-scale colo and power-capex drive outsized revenue growth; small manufacturers, local construction firms and discretionary retailers face demand erosion and margin pressure. Expect leasing spreads to remain elevated in coastal metro cores (lease-rate growth +5–15% YoY over next 12–24 months) while industrial utilization stays flat, concentrating pricing power with a handful of platform players.

Tail risks center on a single-point dependency: if AWS/MSFT/GOOG cut cloud/data-center capex by >15% (a plausible downside scenario on recession or AI ROI re-evaluation) valuation and GDP contributions could roll back sharply; regulatory & permitting moratoria on new data centers or stricter power emissions rules in 12–24 months are credible black swans. Immediate (days–weeks) risks are earnings guidance, short-term (1–3 months) are capex announcements, long-term (6–24 months) are grid constraints and tax/regulatory changes.

Practical trade implications: overweight large-cap data-center REITs and grid operators via equity and 9–15 month call spreads, while underweight industrial cyclicals and small-cap manufacturing suppliers via shorts or inverse ETFs; pair trades (EQIX/DLR long vs XLI short) capture relative strength. Options should be used to cap downside — buy 9–12 month call spreads on DLR/EQIX and buy puts on cyclical industrials around earnings windows.

Contrarian view: consensus overweights headline GDP momentum while underweighting its concentration risk — the market underprices a scenario where power costs compress data-center margins by 200–400bp and hyperscalers delay projects. Historical analogue: late-1990s telecom tower/infra capex boom that reversed; if permitting or capex guidance weakens, expect >25% drawdowns in exposed names, creating deep value entry points.

Key Decisions for Investors

• Establish a 2–3% long position in Digital Realty (DLR) and a 1.5–2% long in Equinix (EQIX) combined (total 4–5% equity allocation) over the next 30 days; set stop-losses at 12% and trim on 20–30% upside or if hyperscaler capex guidance is cut >15% on any quarterly call.
• Initiate a dollar‑neutral pair trade: long DLR (or EQIX) vs short XLI (Industrial Select Sector SPDR) sized 1:1, target 3–9 month horizon, close if manufacturing payrolls exceed +50k/month for two consecutive months or if data-center leasing dips >10% QoQ.
• Buy 9–12 month call spreads on DLR and EQIX (buy ~35-delta calls, sell ~15-delta calls) allocating 0.5–1.0% notional per ticker to cap premium; target >30% return if leasing/utilization stays firm, stop if lease velocity falls 20% vs prior quarter.
• Reduce direct exposure to small-cap industrials and construction suppliers by 30–50% within 2 weeks; redeploy proceeds to utilities with grid exposure (NEE, NRG) via 2–5% combined positions to capture power-demand upside and hedged yield.
• Set a 60–90 day monitoring trigger: add to data-center longs if AMZN/MSFT/GOOG capex guidance remains flat-to-up or if regional power-forward prices rise >10% YoY; cut positions if any hyperscaler reports capex cuts >15% or if state-level moratoria are enacted.