Canada has no shortage of AI ambition. Canada helped invent modern AI, and today more than 150,000 Canadian innovators at over 3,500 companies are building AI solutions in Canada that strengthen our economy, solve real problems, and create good, high paying jobs. SAP is proud to be one of those companies.
Yet as a country we have often failed to turn early leadership into lasting economic advantage and productivity gains. AI for All, the Government of Canada’s new AI strategy, aims to build a responsible, safe, and sovereign AI industry and research community that serves Canadians. The risk is not that we have aimed too low, but that we spread effort so thin across pillars and pilots that nothing moves the productivity dial or consolidates our national AI advantages at the required speed.
Canada has a narrow window to make design choices that will decide whether this strategy delivers real outcomes. AI for All rests on six pillars that seek to protect Canadians, build skills, power shared prosperity, develop a sovereign AI foundation, scale Canadian champions, and deepen trusted partnerships. While these will drive AI adoption across the economy, three levers will matter most for the public sector and regulated industries:
- Clear AI missions with measurable outcomes
- A genuinely sovereign AI stack that keeps sensitive workloads under Canadian control and law
- Procurement rules that prioritize control and let government act as an anchor customer to scale Canadian firms at home and abroad
First, missions. The strategy highlights an AI Missions Program and priority sectors from health and life sciences to manufacturing and public services, but missions cannot be a slogan. They need to be verticals with defined outcomes, while being planned and executed horizontally across central agencies, line departments, and the myriad fragmented procurement processes that actually move money and workloads. An important starting point should be investing in AI-enabled public service delivery: reducing wait times, cutting red tape, and improving frontline service quality are outcomes Canadians will feel. But without a governance mechanism that can align stakeholders across government and industry or a streamlined government IT procurement process, the six pillars risk becoming six different operating logics in practice.
Second, sovereignty and champions. AI for All acknowledges that Canada’s AI value chain spans energy and chips, compute, models, and applications, but that sovereign compute capacity remains nascent, and many critical layers sit under foreign control. Ottawa has responded with a build-partner-buy approach and major investments in AI compute, including a public supercomputer and large-scale data centre projects designed to provide domestically governed capacity. The correct frame for sovereignty is not dogmatic autarky, rather we should aim for thoughtful legal and operational control.
The question of sovereignty and control is not just where data sits, but which jurisdiction’s laws or political whims apply, who operates key infrastructure, and whether Canadians can count on those systems in a crisis. As the strategy puts it, “the underlying infrastructure must be operated under Canadian control and Canadian law,” not on platforms that can be restricted or withdrawn at a foreign government’s discretion. The world learned this lesson last Friday with Anthropic’s announcement of a U.S. government directive to suspend access to the Fable 5 and Mythos 5 frontier AI models.
Third, procurement rules that prioritize control. The government will need to move swiftly to define sovereignty at both a technical and operational level. This means setting standards and ensuring that the definition and standards are reflected across procurement processes. By doing so, digital partners to the Government of Canada can marshal the necessary resources and invest in the infrastructure required to provide sovereign control where needed.
The AI strategy makes it clear that Canada cannot do it alone and will need to find suitable partners and allies to achieve digital sovereignty goals. SAP Canada has established partnerships with Cohere and Bell Canada that aim to bring the best of each company to provide sovereign, security-first cloud options tailored to regulated industries and sensitive government systems. SAP’s expanded partnership with Cohere is designed to deliver sovereign AI solutions by combining Canadian developed foundation models with SAP’s sovereign enterprise platforms. Our alliance with Bell Canada at the infrastructure layer would ensure that our most critical systems and workloads, such as national security systems, utilities, and government services processing sensitive citizen data, remain firmly under Canadian operational control and legal jurisdiction.
The recently announced SAP Sovereign Cloud in France also provides an example of how to enable public bodies and regulated industries to run strategic workloads in environments aligned with the government domestic cloud strategies and stringent national security requirements. This effort will be backed by significant private sector investments in sovereign cloud and trusted Business AI capabilities with infrastructure and AI partners. This demonstrates how strategy-aligned procurement processes can enable long-term partners to help governments bring enterprise cloud and AI capabilities under national legal and operational jurisdiction while preserving choice across multiple trusted platforms and enabling the right level of control for different systems.
AI for All sets forth an ambitious national agenda, from higher business adoption to hundreds of thousands of new jobs and a meaningful lift in Canada’s GDP. Those numbers will only be credible if we can successfully do three things over the next year:
- Name and empower a small number of AI missions with real authority
- Invest now in a sovereign Canadian stack based on a clear definition of sovereignty and rules around jurisdiction and ecosystem access
- Align government procurement rules with anchor customer and alliance principles that will allow government dollars to scale Canadian firms alongside trusted partners
Beneath all this also lies a deeper and more difficult question. Values, ethics, and agency cannot be an afterthought in this new AI era. If Canada wants AI that reflects and preserves our democratic institutions, protects children and workers, respects privacy, and promotes cultural diversity, we must explicitly embed those norms into the laws, contracts, and infrastructure that govern AI. Without these controls, we may find that the systems shaping our economy and democracy are ultimately answerable not to Canadians, but to foreign governments and opaque algorithms.
