Plan unveiled for 'sovereign AI data centre' cluster in Kamloops, Vancouver

Canada and Telus unveiled a three-site 'sovereign AI data centre' cluster in Kamloops and Vancouver, with the Kamloops expansion and first Vancouver facility slated to come online later this year and a second Vancouver site planned for 2029. The project is backed by federal, B.C., and Vancouver officials and emphasizes 98% clean hydro power, 90% less water use than traditional data centres, and waste-heat recovery for up to 150,000 homes. While supportive for Canadian AI infrastructure and Telus, the proposal faces local skepticism over water use, environmental review, and data-centre siting.

Analysis

This is less a single-company headline than a policy signal that B.C. is trying to become the lowest-friction North American jurisdiction for AI infrastructure. The second-order winner is anyone with stranded or underutilized grid access, especially those able to monetize waste heat or secure clean power allocations; the loser is every regional utility and municipality that cannot match the permitting speed, political cover, or “green” narrative. For Telus, the strategic value is not just hosting compute but embedding itself deeper into enterprise AI workflows, which could improve retention in higher-margin digital and security services over the next 12-24 months.

The key market implication is that power and land now matter more than raw networking or cloud branding. If B.C.’s 400 MW allocation becomes a gating mechanism, the scarce asset is access to regulated low-carbon electricity, which should lift the option value of companies with existing interconnects and penalize late entrants forced into higher-cost power or remote builds. Watch for second-order beneficiaries in electrical equipment, cooling, and real-estate conversion plays around Vancouver, while incumbent cloud hyperscalers may face a modest cost disadvantage if sovereign-compute mandates start steering workloads away from U.S.-based platforms.

The main risk is that the headline “green” profile masks local opposition that can stretch timelines from months to years; this is especially true for water-reuse claims and any requirement for new transmission or municipal approvals. The near-term catalyst is political, not financial: any change in provincial leadership, a drought cycle, or a high-profile environmental review could slow the rollout and compress the narrative premium. Conversely, if these sites come online on schedule later this year, it will validate the model and likely accelerate copycat projects elsewhere in Canada.

Consensus may be underestimating how much this benefits domestic incumbents with regulatory fluency relative to pure-play AI infra names. The market wants to own obvious AI compute winners, but the cleaner trade may be the pick-and-shovel layer: grid, cooling, and conversion assets that get paid before utilization risk is fully proven. Telus itself may be a better relative-value story than a directional one: limited upside if the project is successful, but meaningful downside if community or power constraints derail execution.

Key Decisions for Investors

• Long TU on a 6-12 month horizon: asymmetric upside from being a preferred sovereign-compute partner; trim if the stock rerates on announcement-driven enthusiasm before execution risk is cleared.
• Pair trade: long Canadian utility/grid names with B.C. exposure vs short U.S. hyperscaler proxies if sovereign-compute policy expands; thesis is that regulated power access becomes the binding constraint over model demand.
• Buy upside optionality in electrical infrastructure/cooling beneficiaries for 2025-2026 delivery windows; the risk/reward improves if provincial allocation processes tighten further and create a backlog of approved projects.
• Avoid chasing pure AI data-center plays until permitting and water-use scrutiny is clearer; policy support can flip quickly if drought or local opposition becomes the dominant headline.
• If TU underperforms after the initial pop, consider adding on pullbacks rather than momentum-chasing; the better entry is when execution skepticism widens the discount to the strategic value of its grid and real-estate footprint.