Canada declares AI critical infrastructure in bid for mid-power sovereignty
Carney’s plan treats AI on par with energy as Ottawa seeks talent and institutional capacity for a power-hungry buildout.
OTTAWA — Prime Minister Mark Carney released Canada’s new artificial intelligence strategy on Thursday (2026-06-04), pledging to treat AI as critical infrastructure on par with energy and telecommunications. The plan marks an explicit bid to lead the world’s middle powers in the race for sovereign AI capabilities.5
That is a market question because AI’s surging electricity demand is already reshaping power grids and gas-fired generation outlooks. Any government that designates AI as critical infrastructure is pre-committing to build the physical and institutional backbone — data centres, grid connections, talent pipelines — that locks in power demand for years.5
Canada’s roughly 36.5 million residents give it a smaller domestic talent pool than the United States or China, but Carney’s strategy leans on scale relative to peers. Germany, at 82 million, and the UK, France and Italy, all under 100 million, face similar constraints as they jostle for position.1
The talent bottleneck is acute. LinkedIn data show AI added 1.3 million jobs in 2024-25, even as automation displaces workers elsewhere.4 The Atlantic Council defines the gap as one between AI literacy — understanding core capabilities, prompt engineering, evaluating outputs — and AI fluency, the ability to deploy the technology to solve real problems. Canada is betting its education system can close that gap faster than larger rivals.4
Institutional readiness, not headcount, may prove the harder constraint. S&P 500 companies returned over $12 trillion to shareholders through buybacks and dividends between 2015 and 2024, capital that did not build factories or train workers.3 The US industrial base has hollowed out even as its tech giants dominate global markets. Canada is trying to avoid that trap by framing AI infrastructure as a public good akin to energy grids.3
Fragmented global investment rules complicate the buildout. UNCTAD calculates that 63% of global investment flows were subject to a screening regime last year, up from 52% in 2020.2 Industries accounting for 60% of the value of America’s stock markets fall under the potential remit of CFIUS, the US foreign investment committee. Canada’s plan will need to attract foreign capital while navigating those barriers.2
China brings its own advantages. Its population of well over a billion provides a domestic market and talent base no middle power can match, while the United States sits third with just under 325 million residents but dominates AI venture capital and model development.1 Canada is positioning itself as a trusted intermediary, big enough to matter, small enough to avoid the geopolitical drag on US and Chinese tech.5
European skeptics argue the middle-power strategy has limits. Foreign Policy wrote on 2026-06-29 that Europe will never be an AI superpower, warning that if access to advanced models and computing power concentrates in two or three countries, the rest face major economic disadvantages. The EuroStack approach of building European alternatives, the piece argued, does little to address that scenario.6
Canada’s energy endowment gives it one advantage Europe lacks. Cheap hydro power in Quebec and British Columbia, plus emerging small modular reactor projects, can anchor data centre buildout without the grid bottlenecks that plague Germany and other high-cost power markets. The ICE Endex TTF front-month traded at €48.80 and German power futures at €101.34/MWh at the week’s close (2026-07-10), levels that make energy-intensive AI deployment far more expensive in Europe than in Canada.5
Whether institutional capacity can keep pace with policy ambition is still unresolved. Treating AI as critical infrastructure carries real consequences — grid permitting, security classification, supply chain controls — that Canada’s federal system must now implement across provinces with separate energy regulators and labour markets. The strategy’s success will be measured not by the June 4 announcement, but by whether data centre permits clear provincial review within two years and whether Carney can retain the talent to run them.5