Defense procurement can feel slow — sometimes like watching paint dry on the HMCS Halifax. That is by design. National security is, at its core, negotiations span bilateral and multilateral trade considerations, contracts that run into the billions of dollars, and the stakes — from fighter jets to frigates to drones — could not be higher.
Yet speed to decision matters now more than ever. Canada must get ahead not only of a shifting geopolitical landscape, but also of the Fourth Industrial Revolution. Artificial intelligence moves on a three-month cycle, not a three-year one.
Canada has publicly declared sovereign AI infrastructure and compute capacity as a top strategic priority. The Department of National Defense (DND) has stated that its AI strategy will be “foundational to defense modernization,” and has warned that Canada risks losing operational advantage if it falls behind allies and adversaries already accelerating AI adoption. Predictive analytics, surveillance capability, AI-enabled operational systems, and decision-support tools are central to future defense readiness.
As defense systems become increasingly AI dependent, procurement decisions must also become decisions about sovereignty, strategic control and national resilience.
The Munk School of Global Affairs and Public Policy offer a useful frame: sovereignty in the AI era is not isolationism. Canada cannot — and should not — own every layer of the AI stack independently. The goal is preserving “freedom from coercion,” and the ability to act independently under pressure.
That brings us to the real question: who governs this country’s most critical national security and defense data? This tier-one data must remain under Canadian control — free from foreign legal jurisdiction, free from foreign clouds, free from geopolitical coercion.
The challenge runs deeper than geography. Even when data is physically stored in Canada, it may remain exposed if the provider falls under foreign legal jurisdiction, such as the United States CLOUD Act. As the Munk School has argued, data residency does not equal data sovereignty.
Sovereignty is no longer simply about where data sits. It is about whether Canada maintains meaningful control over the systems powering critical national capability, and whether those systems remain resilient if geopolitical, legal or commercial conditions shift.
AI systems are trained on data, shaped by data and dependent on data. In defense environments, that means drones, intelligence platforms, cyber operations, navigation systems, predictive maintenance and military decision-support tools. The DND strategy repeatedly emphasizes that AI capability depends on secure, well-governed and well-architected data environments.
Canada already understands critical minerals as strategic national assets tied to economic and national security. In the AI era, data is the new above-ground critical mineral, the resource feeding the systems increasingly shaping operational readiness and geopolitical leverage.
This does not mean this country should attempt to build every layer of the AI stack independently. Hyperscalers provide capability and scale; interoperability with trusted allies remains essential. But sovereignty still requires Canada to determine where strategic control is non-negotiable.
“Canada First” in the AI era should not mean Canada alone. It should mean Canada is deliberate about which systems, data environments and infrastructure require domestic governance, sovereign protections, or trusted allied arrangements. The challenge is not choosing between total dependence and total independence. It is determining where managed dependency is acceptable and where sovereign control is essential.
What happens when critical national capability depends on infrastructure Canada cannot fully govern, replace, or recover independently? This is where sovereign cloud protections, Canadian-held encryption keys, National Security Exception procurement tools and classified sovereignty thresholds become directly relevant.
The future of defense will be shaped not only by military hardware, but also by the resilience of data environments and AI systems supporting critical decisions.
Canada does not need to own everything. But it must understand what it cannot afford to lose control over.
Procurement must embed data sovereignty questions, criteria, and strategies into its protocols, acknowledging the balance between accelerating decision-making and protecting national security. That balance is achievable, but only if we are deliberate about it.
AI is nothing without data. The defense procurement mandate must answer the real sovereignty question of whether the systems powering this country’s most critical national capabilities will remain governed by Canadian interests, protected under Canadian standards and be resilient when pressure comes.
That is non-negotiable.
