Saint John data centre wouldn't guzzle city water, companies say

A proposed West Saint John data centre would require 390 MW of power, including a 190 MW on-site natural-gas plant, and is expected to emit about 755,187 tonnes of greenhouse gases annually. Water demand appears limited under a closed-loop cooling system, with initial fill estimated at 3.5 million litres, but the project remains controversial due to environmental concerns around wetlands, forest impact, and emissions. If approved, construction could start this year, with operations targeted for 2028 and a 25- to 30-year minimum operating life.

Analysis

The market is likely underpricing the utility- and power-market implications more than the real estate angle. A 390 MW load with a 190 MW on-site gas plant effectively turns this into a quasi-merchant generation project with an AI tenant attached, which is far more relevant for N.B. Power, gas infrastructure, and local transmission planning than the water headline suggests. The second-order issue is that the project may force incremental grid capex and dispatch changes, while also tightening scrutiny on any future hyperscale approvals that lack firm clean-power sourcing.

The biggest beneficiary is not the developer but the upstream energy complex and any firm supplying equipment, turbines, switchgear, gas midstream, and long-duration power engineering services. If the commercial agreement is truly 15-20 years, that improves financing visibility and could de-risk a phased buildout, but it also locks in a carbon-intense load at a time when provincial policy may move toward stricter permitting or emissions offset requirements. That creates a policy overhang: once the project is framed as one of the province’s top emitters, approval odds become a function of local jobs versus climate credibility, not just economics.

The contrarian point is that the water controversy may be a false focal point; the more material constraint is power availability and emissions intensity. If the province or utility delays interconnection, changes tariff structures, or requires lower-carbon supply, economics could compress materially over 6-18 months. Conversely, if approvals proceed, the project could become a template for AI infrastructure in smaller grids, which is bullish for firms that can package behind-the-meter generation and grid services.

Near term, this is a catalyst-heavy setup, not a day trade: EIA review, public consultations, and potential provincial conditions can reprice the project over the next 1-2 quarters. The key tail risk is that higher gas prices or tighter methane/regulatory rules undermine the "cleanest burning" narrative and raise operating costs, while the upside is a multi-year buildout with sticky utility demand and recurring power sales. The best risk/reward likely sits in expressing the theme through power-infrastructure rather than speculative data center equities.

Key Decisions for Investors

• Long utility/gas-infrastructure beneficiaries vs. broad market: express via XLU or regional utility proxies on any pullback over the next 1-3 months; thesis is incremental load and grid-capex demand, with limited earnings sensitivity if the project is delayed.
• Pair trade: long gas infrastructure / power services, short high-duration green infrastructure names over 3-6 months; if approvals lean toward gas-backed data centers, capital will favor reliable dispatch and equipment over pure-play renewables.
• Avoid or underweight pure data-center REIT/AI infrastructure names with weak power visibility; the risk/reward worsens if provincial conditions force cleaner supply or higher interconnection costs.
• Watch for a buying opportunity in N.B. Power-linked/Canadian utility exposures if the market overreacts to emissions headlines; if approved, the incremental load improves utilization and may support rate-base growth over 12-24 months.
• Optionality trade: small long-dated call spreads on power-equipment / electrification names tied to grid buildout, funded by short exposure to carbon-sensitive industrial names if local approval conditions tighten.