Need Energy, Water Solutions With Data Centers: Jan Vesely

Bloomberg Deals discussed climate investing and the infrastructure pressures created by AI data centers, particularly higher energy and water costs. EQT, Pulse Fund and BCG representatives emphasized that data-center operators need to present credible solutions to these resource constraints. The segment is thematic commentary rather than a company-specific financial announcement.

Analysis

The investable implication is not broad "green data center" exposure but a widening dispersion between power-secured operators and capacity-constrained developers. AI workloads convert electricity availability, interconnection rights, and cooling design into scarce assets; operators with contracted low-cost power can defend pricing, while colocators dependent on merchant power face a lagged margin squeeze as utility tariffs and transmission charges reset. This favors infrastructure owners with power-adjacent land and contracted generation over asset-light compute lessors.

Over the next 1-3 months, the relevant catalyst is not sustainability messaging but disclosed power procurement: PPAs, utility service agreements, interconnection queue status, and incremental MW available before 2027. Vertiv (VRT), Eaton (ETN), and GE Vernova (GEV) remain cleaner picks-and-shovels exposure because higher rack density raises spend on power distribution, backup systems, cooling and grid equipment regardless of which cloud tenant wins. The key risk is a hyperscaler capex pause; equipment multiples are vulnerable if 2026 AI capex guidance moderates even while long-run grid demand remains intact.

EQT is not a direct data-center infrastructure proxy despite its institutional-infrastructure adjacency. Treat any thematic read-through as low-conviction unless EQT discloses realizations, fundraising, or portfolio-company economics tied to data-center power and cooling; otherwise the earnings sensitivity is principally management fees, realizations and fundraising conditions. Consensus may also be underestimating water as a siting constraint: areas with cheap power but water restrictions could shift demand toward air-cooled designs, benefiting VRT/ETN but impairing developers whose land banks lack utility and water certainty over 6-18 months.

Key Decisions for Investors

• Maintain a 6-12 month long VRT / short EQIX pair only after confirming VRT order growth remains above 15% and EQIX does not fully pass through utility costs at lease renewal. Thesis: equipment content rises immediately with AI density while data-center margin pressure emerges with a lag; exit if VRT backlog conversion or hyperscaler capex guidance weakens.
• Accumulate ETN on market pullbacks for a 12-18 month grid-and-data-center power buildout exposure; use a 10-12% downside stop or reassess if North American electrical backlog declines for two consecutive quarters. Risk/reward depends on sustained utility and data-center order intake, not ESG capital flows.
• Use GEV as a watch-list long rather than chase: initiate only following evidence that gas-turbine slot pricing and grid-equipment backlog are extending into 2028. The upside is constrained generation supply; falsifier is a material cancellation rate or easing turbine lead times.
• Do not establish a directional EQT position from this discussion alone. Set an event-driven alert around earnings/fundraising disclosures for specific AI-infrastructure AUM, realizations, or fee-related earnings contribution; absent those datapoints, the thematic signal is insufficient.