Alex Davis, CEO of Disruptive, warns that a ‘build it and they will come’ approach to third‑party AI data centers could trigger a significant financing crisis in 2027–2028 as extreme capex and a mismatch between builders and end‑users leave speculative landlords exposed. He contrasts owner‑users (hyperscalers like Microsoft and Meta) that will internalize capacity with REIT‑style landlords (e.g., Digital Realty, Equinix) that could face refinancing pressure and stress private credit markets, even as Davis remains bullish on AI and points to Nvidia’s roughly $20 billion licensing deal for Groq assets as evidence of vertical consolidation. Investors should favor owner‑user exposure and discipline around speculative data‑center developers and lenders to the sector.
Market structure: Chip/IP owners (NVDA, Groq-backed assets) and hyperscaler owner-users (MSFT, META) are the primary winners — they capture margin upside from vertically integrated stacks and can absorb multi-year capex. Speculative wholesale landlords (EQIX, legacy REITs, smaller colocation players, CIFRW-like converts) are exposed to a refinancing cliff in 2027–28 and face demand substitution risk as hyperscalers internalize capacity; expect potential EV/EBITDA multiple compression of ~1–3 turns over 12–24 months for weak landlords. Risk profile: Tail risk is a financing shock in 2027–28 that widens senior unsecured spreads for data-center REITs by 200–500bps and forces asset sales; secondary tails include power-grid shortages and anti-competitive regulatory scrutiny of chip/IP consolidation. Near-term (days–weeks) volatility should be modest; medium-term (3–12 months) is driven by guidance and capex; long-term (2027–28) is the refinancing/capacity inflection. Hidden dependencies include customer contract tenure (multi-year vs month-to-month) and concentrated capex exposure to specific hyperscalers. Trade implications: Favor long concentrated AI/IP positions (NVDA) and owner-user cloud equities (MSFT/META) while selectively shorting landlords with high leverage and customer concentration (EQIX, CIFRW, DLR) into any rallies. Use duration-limited options to express bearish views on landlords (9–18 month puts or put spreads) and buy credit protection on 2027–2029 maturities if spreads widen >200bps. Rotate away from REIT beta into semis/cloud infra over next 3–12 months. Contrarian angles: Market may underappreciate differentiated landlords (Equinix) that provide irreplaceable interconnect/ecosystem value — these could trade as strategic assets not commoditized capacity. Alternatively, a disorderly capex pullback by hyperscalers could create a supply squeeze and a short-covering rally in 2028; define stop-losses and scale positions so a >30% move against shorts triggers re-evaluation. Historical parallel: telecom fiber rollouts saw consolidation then re-rating; similar outcome possible here for high-quality landlords.
AI-powered research, real-time alerts, and portfolio analytics for institutional investors.
Request a DemoOverall Sentiment
moderately negative
Sentiment Score
-0.35
Ticker Sentiment