Submission to the House of Commons Standing Committee on Finance
Pre-Budget Consultations in advance of the Fall 2025 Federal Budget
Submitted on behalf of Canada’s leading Hardtech incubator, ventureLAB, whose mission is to build and scale globally competitive ventures that advance Canada's knowledge-based economy.
Executive Summary
Canada has a storied history of excellence in artificial intelligence (AI) research. Canada’s AI leadership presents significant potential to drive economic growth, yet startups face considerable challenges in transforming cutting-edge AI technologies into globally competitive products, resulting in missed opportunities in improving Canada’s productivity using AI, and in establishing Canada’s global leadership in commercializing and scaling deeptech innovations.
To seize this strategic opportunity in Budget 2025, we recommend that the federal government support three national priorities developed by leading industry and ecosystem voices:
- Create a “Buy Canadian” AI Commercialization Strategy to accelerate the adoption of domestically developed AI solutions, especially those developed by Canadian-based small and medium-sized enterprises (SMEs).
- Establish a Pan-Canadian AI Compute Centre for SMEs, consisting of energy-efficient (“greener AI”) small-scale or micro data centres (MDCs) that include a mix of industrial-strength and experimental, domestically developed infrastructure technologies for scaling Canadian AI-powered SME solutions. The proposed centre will serve not only as a sandbox for testing Canadian AI-powered solutions, but it also offers a platform for adopting developers of hardware infrastructure technologies (semiconductors, optical processor interconnects, etc.).
- Upskill the workforce for Responsible AI, creating a resilient workforce that is prepared for and can drive AI transformation in Canada, with a focus on adapting and adopting Canadian-based solutions that are backed by solid foundational knowledge of responsible AI best practices.
Background
Canada has long been recognized as a cradle of AI innovation. Canadian-based AI researchers from coast to coast have been conducting seminal research on neural networks and reinforcement learning.
In 2017, Canada became the first nation in the world to launch a national AI strategy, the Pan‑Canadian Artificial Intelligence Strategy, backed by approximately CAD $125 million in funding to build research capacity and support institutes like Mila (Montreal), the Vector Institute (Toronto), and Amii (Edmonton). This strategy has since expanded, with combined investments reaching over CAD $2 billion toward AI and digital research and innovation since its inception.
Canada’s AI ecosystem is powered by world-class talent and a research ecosystem fuels it.
- Thought leadership and global reputation: The scholarship of leading AI researchers, most notably Geoffrey Hinton, Yoshua Bengio, and Richard Sutton, underscores Canada’s pivotal role in developing deep learning and reinforcement learning frameworks.
- Research infrastructure and collaboration: Institutions under the Pan‑Canadian AI Strategy, Mila, Vector and Amii serve as hubs that attract top talent and foster collaboration among academia, government, and industry.
- Ethical and policy leadership: Canada’s early emphasis on responsible AI, including ethical frameworks and safety dialogues, sets a benchmark globally.
Despite these strengths, Canada faces significant challenges in commercializing and scaling domestically developed AI innovations:
- Access to affordable AI Compute by Canadian-based AI SME: AI SMEs typically operate on limited budgets, often relying heavily on cloud services (e.g., AWS, Google Cloud) that charge by usage, which is financially unsustainable when training large models. While researchers may have access to some degree of shared computing infrastructure (like university supercomputing clusters), startups often face higher barriers due to the ongoing operational costs of AI compute (e.g., GPUs, TPUs, storage). The lack of localized access to affordable AI compute resources leads to constraints on the ability of startups to experiment and scale models.
- Sovereign AI infrastructure and access to domestically developed AI infrastructure: Canada faces unique challenges when it comes to controlling and managing the AI Compute resources within its borders, including hardware, software and other compute infrastructure components. By relying on foreign cloud services, Canadian AI SMEs have limited control over their AI compute resources and may face challenges regarding data residency and cross-border data flows. For example, U.S.-based providers must comply with U.S. laws, including the Cloud Act, which allows U.S. government agencies to access data stored by companies, even if that data is stored outside of the U.S. While Canada has world-class AI research institutions, there is a lack of dedicated national AI infrastructure that provides AI SMEs with sufficient access to advanced AI Compute resources that include domestically developed infrastructure solutions.
- AI adoption lag: Small- and medium-sized enterprises are Canada’s economic backbone. They often lack awareness, capital, or infrastructure to experiment with AI. As a result, they remain underrepresented among AI adopters. While Canada excels in research, many promising breakthroughs are commercialized abroad, drawing both investment and talent away from Canada.
- Workforce readiness gaps: Canadian business use of AI is modest, especially among SMEs. StatsCan found just 6.1% of businesses were using AI in Q2-2024. In a separate survey, only 38.9% of the AI adopters trained staff to use AI, signalling a significant upskilling gap. Canadian labour studies show fast-rising demand for AI skills from 2018–2024, but responsible AI capabilities such as governance, privacy security, and impact assessment lag the technical track. Even when there is an appetite to pursue upskilling initiatives, the organization undertaking those initiatives face time and budgetary constraints, resulting in a limited number of employees being trained. In short, there are gaps in upskilling the workforce at scale and at optimal efficiency.
The Opportunity
Massive productivity gains projected
Canada possesses deep AI talent and leadership in research, but is held back by insufficient infrastructure that helps scale Canadian-developed AI solutions, a comprehensive plan for adopting Canadian-developed AI solutions, and an overall strategy for upskilling the entire workforce for responsible AI development and adoption.
Accenture Canada estimates that AI could deliver approximately CAD $180 billion in annual productivity gains by 2030. Additionally, The Conference Board of Canada predicts AI could raise national labour productivity by over 17%, helping reverse Canada’s long-standing productivity declines.
Recommendations
- Create a “Buy Canadian” AI Commercialization Strategy, with policies and incentives for both government and the private sector, aimed at accelerating the adoption of domestically developed AI solutions, especially those developed by Canadian-based small and medium-sized enterprises (SMEs). The low adoption rate of AI by Canadian businesses suggests that there is an opportunity to create a strategically designed national initiative that fosters the adoption of AI, resulting in overall productivity gains. Targeted procurement and adoption strategies can pull innovations from labs into production, and from domestically developed AI solutions by AI SMEs to national rollout of these solutions. Various incentives, trade-compliant procurement, policy adjustments, and transparent KPIs can help kickstart the initiative, and forge a culture of “buying local” among the Canadian businesses. An important component of the strategy must include responsible AI best practices, ensuring that Canadian-produced AI solutions adhere to the highest ethical standards, thereby creating a differentiated value proposition for Canadian AI solutions.
- Establish a Pan-Canadian AI Compute Centre for SMEs, consisting of a federated network of small-scale or micro data centres (MDCs) that doubles as a sandbox for Canadian AI solutions and a platform for domestic hardware innovators (chips, photonics, interconnects). The federal government has already launched a national Sovereign AI Compute Strategy and opened an AI Compute Access Fund to give SMEs affordable compute. While the existing Sovereign AI Compute Strategy and AI Compute Access Fund represent important first steps, they do not fully address the underlying infrastructure challenges facing Canadian SMEs. Current programs don’t address SME need for distributed access to regional compute infrastructure and could still leave SMEs dependent on foreign cloud providers, and don’t create cost effective opportunities to test and scale domestically developed hardware solutions. The proposed centre complements these investments, operationalizing in-region access on fair terms, while catalyzing the adoption of domestically developed infrastructure solutions where applicable. Canada has experience establishing such sandboxes in other sectors, such as networking/5G testbeds (CENGN, ENCQOR). Canadian-based SMEs have used shared, production-like infrastructure to validate products and scale. The proposed centre brings the same model to AI compute. Further, building these MDCs enables Canada to adopt energy-efficient, sustainable solutions where applicable. International peers are already moving in this direction. South Korea’s Ministry of Science and ICT plans to build micro data centers (MDCs) based on locally manufactured AI chips to serve Korean SMEs, while expanding national AI chip production, driving deployments of AI services, and supporting the national SME sector. Canada needs a made-in-Canada variant.
Beyond supporting SME growth, this infrastructure investment is critical to Canada’s economic security. As AI becomes the defining technology of the 21st century, nations without sovereign AI compute capabilities risk becoming dependent on foreign powers for their most strategic economic assets. By continuing to build domestic AI infrastructure now, Canada would secure its ability to control its data, protect its innovations, and ensure that Canadian businesses can compete globally without relying on infrastructure controlled by other nations.
- Upskill the workforce for Responsible AI development and adoption, creating a resilient workforce that is prepared for and can drive AI transformation in Canada, with a focus on adapting and adopting Canadian-based solutions that are backed by solid foundational knowledge of responsible AI best practices. In short, there should be a general understanding of the compliance context for real deployments, such as European Union's AI Act, ISO 42001 Compliance, and Québec’s Law 25 (ADM transparency & PIAs). This general knowledge should be part of “day-one” literacy for builders, buyers, and users of AI. Further, the Canadian workforce should have an understanding of AI safety and risk management considerations, addressing issues such as trust, data governance, security vulnerabilities, and robustness, alongside ethical concerns regarding privacy and fairness.
Conclusion
Adopting AI in Canada, especially domestically developed AI solutions, is a strategic imperative, and can be a powerful catalyst for driving productivity gains, propelling economic growth, improving efficiency for businesses of all sizes, empowering workers, and driving long-term competitiveness and economic security.
We recommend proceeding with these transformative national projects to adopt domestically developed AI solutions, develop a Pan-Canadian AI Compute Centre for SMEs, and upskill the workforce for responsible AI.
We thank you for your consideration and welcome the opportunity to support Canada’s leadership in this critical industry.
For more information, please contact: sdodds@venturelab.ca
