Nscale swallows lion's share of UK datacenter investment

Nscale has raised $3.7 billion, accounting for nearly two-thirds of the $5.7 billion in equity funding raised by UK datacenter infrastructure companies, and is the sector's only unicorn, according to Tracxn. Funding has accelerated with demand for AI infrastructure, while London captured 93% of tracked investment and nearly 100 UK datacenter construction projects worth more than £36 billion ($49 billion) are in the pipeline. Capacity and infrastructure execution remain risks: 43% of surveyed operators reported needing upgrades or remedial work after facilities became operational.

Analysis

The relevant public-market implication is not UK private-funding momentum itself, but a potential bottleneck shift from GPU availability toward powered, commissioned capacity. If retrofit and remediation requirements remain elevated, operators face delayed revenue commencement and capex leakage; this favors scaled platforms EQIX and DLR, whose procurement, operating expertise, and customer cross-connect ecosystems can absorb execution risk better than smaller colocation entrants. The second-order beneficiary is the hyperscaler: MSFT and GOOG can use long-duration commitments and balance-sheet capacity to secure scarce high-quality capacity, reinforcing barriers for smaller AI-cloud providers.

Near term (days to 1 month), this is unlikely to move EQIX or DLR absent evidence of incremental signed leases, power availability, or revised development yields. Over 1-3 months, watch management commentary on AI-related interconnection demand, preleasing, construction completion dates, and stabilized yields; a pattern of completion slippage alongside sustained bookings would support pricing power but penalize operators with uncontracted development exposure. Over 6-18 months, a shortage of reliable powered capacity should expand the strategic value of incumbent campuses, but only if power costs and grid-connection delays do not push development returns below underwriting assumptions.

Contrarian view: investors may over-credit data-center landlords for AI demand while underestimating that hyperscalers retain bargaining power through self-build, geographic diversification, and proprietary infrastructure. EQIX is less a pure GPU-capacity beta than DLR because its network-dense retail colocation mix is comparatively defensive; that reduces downside but also limits direct upside from wholesale AI deployments. The thesis is falsified if hyperscaler capex remains high while third-party leasing, pricing, and utilization fail to improve—indicating AI capacity is being internalized rather than monetized by listed operators.

Key Decisions for Investors

• Maintain a 6-12 month relative long DLR / short EQIX only if upcoming results show DLR accelerating hyperscale preleasing and development yield while EQIX's AI-related bookings remain concentrated in interconnection. Target a 10-15% relative return; exit if DLR development completion dates slip materially or EQIX raises guidance on recurring interconnection revenue.
• Do not chase either REIT on this report alone. Set alerts for UK/European power-connection delays, commissioning slippage, and any revision to construction budgets: these are the highest-frequency indicators that nominal AI demand will not translate into near-term rentable capacity.
• Use MSFT and GOOG as higher-quality 6-18 month AI infrastructure exposures rather than direct UK data-center bets. Their ability to contract capacity and fund self-build makes them relative winners if third-party capacity tightness persists; reassess if cloud capex guidance decelerates or AI revenue monetization misses begin to constrain spend.
• For a more tactical infrastructure expression, consider initiating DLR only on evidence of incremental AI leasing rather than sector funding headlines. Require disclosed preleasing or improved stabilized-yield guidance; without it, the risk/reward is vulnerable to rate-driven multiple compression and elevated construction capex.