This sector has been adopting AI at a "surprising" rate, Barclays says

Barclays strategists report rapid AI adoption in real estate—citing JLL that 88% of investors are piloting AI and Eurostat data showing the sector has the third-highest AI usage in Europe—and estimate AI could cut costs, boost margins and add up to 15 percentage points of per-share profit growth in a 10-year 'blue-sky' scenario. McKinsey projects $110–$180 billion of additional value for the sector, while Barclays sees medium- to long-term upside for revenue generation, flags mixed sub-sector impacts (non-prime offices vulnerable, data centers advantaged), and assigns overweight ratings to Merlin Properties and Tritax Big Box as key beneficiaries of data-center expansion.

Analysis

Market structure: Winners are data-center-focused REITs and scale players that can monetize AI (Merlin Properties MRL, Tritax Big Box BBOX.L, server vendor SMCI, AI-driven ad/software names like APP) because AI lifts demand for specialized real estate and compute, potentially adding 15 percentage points to EPS growth in a 10-year blue-sky and $110–180bn to sector value (McKinsey). Losers include non-prime office landlords and small-cap regional REITs that lack capex firepower; expect pricing power to concentrate with large landlords and hyperscalers, pressuring smaller owners' occupancy and rents.

Risk assessment: Tail risks include regulatory limits on data flows/privacy or an EU/UK moratorium on new large data facilities, a sharp commodity/energy spike (>20% electricity cost rise) or a financing shock (10yr Treasury +75bp in 3 months) that swings REIT cap rates and wipes 10–30% off valuations. Time horizons: immediate (days–weeks) will be driven by capital flows and M&A chatter; short-term (1–6 months) by earnings and capex announcements; long-term (3–5 years) by realized AI-driven revenue uplift and capex payback curves.

Trade implications: Direct plays: establish a 2–3% long position in MRL and 1–2% in BBOX.L as primary data‑center real estate exposure, and a 1% tactical long in SMCI for hardware upside. Pair trade: long MRL vs short suburban/non-prime office REIT OFC (equal dollar exposure) to capture dispersion; options: buy 9–12 month SMCI call spreads (e.g., 1x 0.5–1.0 delta long call financed by 0.2–0.3 delta calls) and buy 6‑month 5–7% OTM puts on MRL as rate/permit shock protection.

Contrarian angles: Consensus underestimates capex and energy headwinds — adoption is likely concentrated among large landlords, raising risk of data-center oversupply in select markets and increasing local regulatory friction that could delay projects by 12–24 months. If power/permit issues materialize, expect localized valuation hits of 20–40%, so prefer assets with contracted cashflows to speculative greenfield builds.

Key Decisions for Investors

• Establish a 2–3% portfolio long in Merlin Properties (MRL) over the next 2–6 weeks, target entry near a 5–10% pullback from recent highs; set a stop loss at -15% and take profit in tranches at +25% and +50% over 12–24 months, re-evaluate if 10yr Treasury rises >75bp.
• Build a 1–2% position in Tritax Big Box (BBOX.L) within 1 month to capture European data‑centre expansion; hedge with a 6–12 month put (5–10% OTM) if local permitting risk increases or electricity costs spike >15% YoY.
• Initiate a 1% tactical long in Super Micro Computer (SMCI) funded by selling a shorter-dated call (buy 9–12 month 0.5–0.7 delta call, sell 0.2–0.3 delta call) to express upside while limiting cost; close or roll if gross margins compress by >300bp on next two quarters' results.
• Implement a pair trade: go long MRL and short OFC (Corporate Office Properties Trust) equal dollar size (1–2% each) to play AI benefiting data-centres vs non-prime office secular decline; rebalance if OFC occupancy stabilizes >70% for two consecutive quarters.
• Reduce exposure to small-cap regional office REITs by 40–60% within 30 days and redeploy proceeds into scale data-center REITs or SMCI, unless those REITs secure multi-year NN lease contracts covering >70% of revenue.