By Chris Hogg
DIGITAL JOURNAL
March 12, 2026
Image by Gemini AI / Google
Canada played a central role in the breakthroughs behind modern artificial intelligence. Now some experts warn the country risks becoming a customer in the very industry it helped create.
A new report titled Sovereign by Design: Strategic Options for Canadian AI Sovereignty argues that Canada risks becoming dependent on foreign artificial intelligence systems unless it takes a more deliberate approach to building and governing the infrastructure behind the technology.
For anyone trying to understand the rapid shifts happening across the AI sector, the report offers a clear explanation of what’s actually at stake and moves the conversation beyond research breakthroughs to focus instead on the infrastructure that determines who ultimately controls and benefits from artificial intelligence.
The authors lay out why this question is increasingly tied to economic security and long-term competitiveness, and they argue the next several years will determine whether Canada remains an active participant in shaping the AI economy or becomes largely dependent on systems developed and controlled elsewhere.
The report was authored by Jaxson Khan and Sean Mullin, senior fellows at the Munk School of Global Affairs & Public Policy at the University of Toronto.
Their analysis, published through the AI Competitiveness Project, examines how the global AI economy is increasingly being shaped by the infrastructure required to deploy artificial intelligence at scale.
Khan previously served as a policy advisor to Canada’s minister of innovation and helped shape the federal government’s sovereign AI compute strategy. Mullin is an economist who advised the prime minister’s office on economic policy and has worked extensively on Canada’s innovation and industrial strategy.
The authors characterize artificial intelligence as a foundational technology that will influence how industries evolve and how economic value is distributed.
Canada helped pioneer the scientific breakthroughs behind modern machine learning. But the systems that will power the next phase of the AI economy are increasingly being built elsewhere.
Sovereignty in the AI era
The report, released this week, focuses on a concept that’s becoming central in technology policy debates: AI sovereignty.
Khan and Mullin describe sovereignty as the ability for a country to make independent economic and political decisions without being subject to coercion from external technology systems or foreign infrastructure providers.
In that sense, sovereignty means maintaining influence over the capabilities that shape how AI systems are built, deployed, and governed.
The report defines the AI ecosystem as a layered stack that includes physical computing infrastructure, cloud platforms, the models themselves, and the software platforms used to deploy those systems across industries.
Control of those layers influences where companies build products, where talent clusters, and where economic value accumulates.
By looking at the stack this way, Khan and Mullin identify specific “chokepoints” where Canada is most exposed. While the country is strong in research and governance, it is significantly underweight in the hardware and cloud layers that act as the gatekeepers for the rest of the stack.
The infrastructure chokepoint
The authors also say artificial intelligence should be understood as a general-purpose technology similar to electricity or the internet. These technologies don’t simply create new products — they reshape entire economies and become foundational infrastructure that other industries rely on.
Artificial intelligence is now entering that phase.
But the report argues Canada’s current policy frameworks often treat very different AI systems and datasets as if they require the same level of control.
Khan and Mullin argue that a more practical approach is to treat AI infrastructure in layers, with different levels of sovereignty depending on how sensitive the underlying data and systems are.
A layered approach to data
At the most restrictive level are systems tied to national security or critical government operations, which the authors argue should run on infrastructure fully controlled by Canadian institutions.
Other systems can operate on domestic cloud platforms governed exclusively by Canadian law. In less sensitive cases, organizations may still rely on global cloud providers, provided legal and contractual protections ensure that Canadian data and systems remain under Canadian jurisdiction.
Applying this kind of layered sovereignty framework would require changes to how the Canadian government handles security classification.
Khan and Mullin point out that current standards are often too rigid, which leads to over-classification where non-sensitive data is treated with the same extreme caution as national security secrets. This creates a bottleneck that prevents government agencies from using the most advanced AI tools.
To make a sovereign strategy work, the report suggests that Canada must modernize its classification rules to clearly distinguish between data that requires a domestic “self-hosted” cloud and data that can safely run on global platforms with the right legal protections. Without this clarity, the country risks overpaying for security on one hand or leaving critical assets vulnerable on the other.
Unlike previous industrial revolutions that unfolded over decades, the AI economy is scaling at an unprecedented speed.
Khan and Mullin argue that the infrastructure decisions being made now will shape how the industry operates for decades. Once companies and governments build their operations on particular cloud platforms and computing environments, switching becomes expensive and disruptive.
If Canada waits too long to establish its own infrastructure and governance frameworks, the authors warn the country could end up deeply dependent on systems controlled elsewhere.
Much of the infrastructure supporting AI development is concentrated within a small number of global technology companies that operate the cloud platforms and computing environments required to train and deploy advanced systems.
Canada, by contrast, has historically focused on research and early-stage innovation. That strategy helped establish world-leading AI institutes and produced influential researchers whose work helped define the field.
But the report suggests research leadership alone doesn’t guarantee economic leadership.
The authors point to a recurring Canadian pattern where the country generates important scientific breakthroughs but allows the industrial and commercial value to be captured elsewhere.
This disconnect is visible in Canada’s current AI ecosystem. While the country possesses world-class research institutions, promising startups, and expanding compute investments, these elements are not yet working in tandem. There is still no unified industrial strategy to align these pieces for building and operating AI systems at scale.
A significant part of this alignment problem stems from what the report describes as fragmented machinery of government.
Currently, the policies governing Canada’s digital and AI infrastructure are spread across multiple federal departments, which prevents the country from speaking with a single, unified voice.
Khan and Mullin argue that this lack of coordination makes it difficult to negotiate with global technology giants or to manage large-scale infrastructure projects.
To solve this, the authors suggest that Canada needs a more centralized authority or a dedicated digital agency to oversee the “sovereign by design” framework and ensure that investments in compute and cloud services are actually meeting the country’s strategic goals.
Where the AI economy is actually controlled
Much of the public conversation about artificial intelligence focuses on the latest models and research breakthroughs. But Khan and Mullin argue that the real leverage in the AI economy sits deeper in the technology stack.
The systems that train, host, and distribute artificial intelligence are increasingly determining where companies build products, where talent gathers, and where economic value accumulates.
Artificial intelligence systems depend on a stack of technologies that includes computing infrastructure, cloud hosting environments, the models themselves, and the applications that run on top of them.
Each layer concentrates power in different ways.
At the base of the stack is compute, the specialized chips and massive data centres required to train and run advanced AI systems. These facilities require enormous capital investment and long-term operational capacity.
Above that sits cloud infrastructure, which allows companies to access computing power and AI capabilities without owning their own data centres.
On top of the cloud sit the models themselves, the systems trained to generate text, images, predictions, or software code.
And finally come applications, where companies integrate artificial intelligence into products, services, and internal workflows.
Most public attention focuses on the models, but much of the economic power sits in the infrastructure layers underneath them.
Control the infrastructure and companies build their products on your systems. Talent gathers around those platforms. Data flows through those networks. Economic value accumulates within that ecosystem.
The report says the consequences of this are already visible in the global AI economy.
A small group of technology firms now controls much of the infrastructure required to train and deploy advanced AI systems. Their cloud platforms host the tools that thousands of companies rely on to build AI-driven products.
For countries like Canada, that concentration creates a strategic dilemma.
Canada has produced world-class research talent and influential AI institutions. Yet the infrastructure layer where much of the economic value is captured is largely controlled by companies headquartered elsewhere.
The report notes that Canada’s vulnerability is especially bad at the model layer.
While the global market is dominated by a handful of American giants, Canada possesses a unique strategic asset in Cohere, which remains one of the only world-class commercial foundation model companies based outside the United States. Khan and Mullin argue that supporting companies like Cohere is vital for maintaining a sovereign alternative.
Additionally, the authors highlight the growing importance of open-source and open-weight AI models. These systems provide a way for Canadian firms to avoid being locked into the proprietary “black box” ecosystems of foreign providers, which allows for greater transparency and local control over how AI is deployed.
Sovereignty in the AI era isn’t about building a national champion. It’s about making sure domestic firms retain meaningful access to the infrastructure that will shape their industries.
Canada’s strategic choices
Khan and Mullin argue that Canada’s strategy doesn’t hinge on building every layer of the AI stack domestically.
Instead, they point to a handful of areas where policy choices could influence how Canada participates in the global AI economy. One of the most immediate is computing infrastructure.
Training and deploying advanced AI systems requires enormous computing power. Countries that host large-scale compute infrastructure often become hubs for AI research, startup formation, and enterprise adoption.
Canada has already begun investing in this capacity.
The federal government allocated $2.4 billion in the April 2024 federal budget to expand sovereign AI compute infrastructure and ensure Canadian researchers and companies have access to the computing resources required to develop advanced systems.
The leverage: Electricity and minerals
While building domestic capacity is a priority, the report argues that Canada should also use its unique natural advantages to secure its place in the global market. Canada possesses two critical assets that global AI firms desperately need: an abundant supply of clean, low-carbon electricity and a wealth of critical minerals required for hardware production.
Khan and Mullin suggest that Canada can use these resources as strategic leverage. Instead of acting as a passive host for foreign data centres, the country could trade access to its energy grid for guaranteed access to high-end hardware or a commitment to keep certain data and intellectual property within Canadian borders. This would transition Canada from a typical customer to a strategic partner that possesses essential resources.
Beyond direct investment, the report highlights the critical role of government procurement. As one of the largest buyers of technology in the country, the federal government has the power to act as a “first customer” for domestic AI firms.
Currently, much of that spending flows to foreign platforms, which further entrenches Canadian dependence on outside infrastructure. By intentionally directing procurement toward sovereign Canadian providers, the government can help these firms reach the scale they need to compete internationally. This approach would turn public spending into a tool for industrial strategy, ensuring that the tax dollars used to modernize government services also help build Canada’s domestic AI capacity.
Infrastructure is only part of the equation.
The bigger economic impact comes when artificial intelligence is embedded inside major industries such as manufacturing, energy production, financial services, agriculture, and logistics.
Canada’s economy includes globally competitive companies in many of these sectors.
That creates an opportunity for Canadian firms to become leaders in applying artificial intelligence to complex industrial systems rather than competing directly in the global race to build foundational AI models.
The report also emphasizes the importance of international alliances.
Artificial intelligence is inherently global. Few countries can build every element of the technology stack on their own.
Through coordinated investments, shared infrastructure, and aligned governance frameworks, those countries could collectively strengthen their position in the global AI ecosystem and reduce dependence on a small number of dominant technology platforms.
Khan and Mullin argue that by aligning with countries like the United Kingdom, France, and Japan, Canada can move beyond its bilateral relationship with the United States.
These nations face similar structural challenges because they all possess world-class research talent but lack the massive scale of the American tech giants. The authors suggest that a coordinated alliance would allow these middle powers to pool their resources, share computing power, and establish common standards for data privacy and security.
This collective approach gives Canada a seat at the table when global rules for artificial intelligence are being written, which ensures that the country is not simply forced to adopt frameworks designed elsewhere.
Khan and Mullin argue that sovereignty doesn’t mean trying to build every piece of the AI stack domestically. For a country the size of Canada, that wouldn’t be realistic.
Instead, they suggest ensuring Canada maintains enough domestic capability and alternative options that it isn’t locked into a single foreign platform. Supporting sovereign cloud providers, encouraging open-weight models, and maintaining access to multiple infrastructure partners would give Canadian firms room to move if conditions change.
That flexibility matters if prices rise, rules change, or geopolitical pressures reshape how global technology platforms operate.
The report emphasizes that Canada can’t outspend superpowers like the United States or China. While those nations can commit hundreds of billions of dollars to achieve AI dominance, Canada must instead rely on being smarter about how it structures its dependencies.
A central part of this strategy involves preparing for the July 2026 review of the Canada-United States-Mexico Agreement (CUSMA). The authors warn that this review presents a significant risk to digital sovereignty, as Canada may face intense pressure to trade away its ability to regulate its own digital infrastructure in exchange for traditional trade concessions.
To protect its long-term interests, the report argues that Canada must treat its AI policy as a non-negotiable part of its national security during these talks.
What this means for business leaders
For executives, the issue often shows up as a technology decision, but the report argues the implications run much deeper.
Artificial intelligence is rapidly becoming embedded in the core operations of many businesses. Financial institutions use machine learning to detect fraud. Retailers analyze purchasing patterns to refine pricing strategies. Manufacturers rely on predictive models to manage supply chains.
These tools promise efficiency and competitive advantage.
But they also introduce new forms of operational dependency.
When a company runs critical systems on infrastructure it doesn’t control, it becomes subject to the pricing models, policies, and legal frameworks of the platforms providing that infrastructure.
If a Canadian manufacturer, bank, or logistics firm builds core operational capabilities on foreign cloud systems, those systems effectively become part of the country’s economic infrastructure. That also means those operations may ultimately fall under the laws and jurisdiction of foreign governments, something the report notes is already a concern in cases such as the United States’ CLOUD Act.
To understand this risk, the report distinguishes between data residency and data sovereignty.
Data residency simply refers to where the servers are physically located. Many foreign cloud providers have built data centres in Canada to satisfy residency requirements, but Khan and Mullin argue this is not enough.
Data sovereignty, by contrast, refers to which country has the legal authority over that data. Under laws like the U.S. CLOUD Act, the American government can sometimes compel U.S. companies to provide access to data even if it is stored on Canadian soil.
For Canadian businesses and government agencies, this means that physical location does not always guarantee legal protection. True sovereignty requires using providers that are subject only to Canadian jurisdiction.
That raises strategic questions for business leaders:Where are your AI workloads running?
Who governs the platforms processing your data?
What happens if those systems change their pricing, access rules, or regulatory obligations?
These questions go beyond technology strategy. They touch on economic resilience and long-term competitiveness.
The next question is whether the country will help build the systems that run on top of those breakthroughs, or whether Canadian companies will simply plug into platforms developed somewhere else.
March 12, 2026
Image by Gemini AI / Google
Canada played a central role in the breakthroughs behind modern artificial intelligence. Now some experts warn the country risks becoming a customer in the very industry it helped create.
A new report titled Sovereign by Design: Strategic Options for Canadian AI Sovereignty argues that Canada risks becoming dependent on foreign artificial intelligence systems unless it takes a more deliberate approach to building and governing the infrastructure behind the technology.
For anyone trying to understand the rapid shifts happening across the AI sector, the report offers a clear explanation of what’s actually at stake and moves the conversation beyond research breakthroughs to focus instead on the infrastructure that determines who ultimately controls and benefits from artificial intelligence.
The authors lay out why this question is increasingly tied to economic security and long-term competitiveness, and they argue the next several years will determine whether Canada remains an active participant in shaping the AI economy or becomes largely dependent on systems developed and controlled elsewhere.
The report was authored by Jaxson Khan and Sean Mullin, senior fellows at the Munk School of Global Affairs & Public Policy at the University of Toronto.
Their analysis, published through the AI Competitiveness Project, examines how the global AI economy is increasingly being shaped by the infrastructure required to deploy artificial intelligence at scale.
Khan previously served as a policy advisor to Canada’s minister of innovation and helped shape the federal government’s sovereign AI compute strategy. Mullin is an economist who advised the prime minister’s office on economic policy and has worked extensively on Canada’s innovation and industrial strategy.
The authors characterize artificial intelligence as a foundational technology that will influence how industries evolve and how economic value is distributed.
Canada helped pioneer the scientific breakthroughs behind modern machine learning. But the systems that will power the next phase of the AI economy are increasingly being built elsewhere.
Sovereignty in the AI era
The report, released this week, focuses on a concept that’s becoming central in technology policy debates: AI sovereignty.
Khan and Mullin describe sovereignty as the ability for a country to make independent economic and political decisions without being subject to coercion from external technology systems or foreign infrastructure providers.
In that sense, sovereignty means maintaining influence over the capabilities that shape how AI systems are built, deployed, and governed.
The report defines the AI ecosystem as a layered stack that includes physical computing infrastructure, cloud platforms, the models themselves, and the software platforms used to deploy those systems across industries.
Control of those layers influences where companies build products, where talent clusters, and where economic value accumulates.
By looking at the stack this way, Khan and Mullin identify specific “chokepoints” where Canada is most exposed. While the country is strong in research and governance, it is significantly underweight in the hardware and cloud layers that act as the gatekeepers for the rest of the stack.
The infrastructure chokepoint
The authors also say artificial intelligence should be understood as a general-purpose technology similar to electricity or the internet. These technologies don’t simply create new products — they reshape entire economies and become foundational infrastructure that other industries rely on.
Artificial intelligence is now entering that phase.
But the report argues Canada’s current policy frameworks often treat very different AI systems and datasets as if they require the same level of control.
Khan and Mullin argue that a more practical approach is to treat AI infrastructure in layers, with different levels of sovereignty depending on how sensitive the underlying data and systems are.
A layered approach to data
At the most restrictive level are systems tied to national security or critical government operations, which the authors argue should run on infrastructure fully controlled by Canadian institutions.
Other systems can operate on domestic cloud platforms governed exclusively by Canadian law. In less sensitive cases, organizations may still rely on global cloud providers, provided legal and contractual protections ensure that Canadian data and systems remain under Canadian jurisdiction.
Applying this kind of layered sovereignty framework would require changes to how the Canadian government handles security classification.
Khan and Mullin point out that current standards are often too rigid, which leads to over-classification where non-sensitive data is treated with the same extreme caution as national security secrets. This creates a bottleneck that prevents government agencies from using the most advanced AI tools.
To make a sovereign strategy work, the report suggests that Canada must modernize its classification rules to clearly distinguish between data that requires a domestic “self-hosted” cloud and data that can safely run on global platforms with the right legal protections. Without this clarity, the country risks overpaying for security on one hand or leaving critical assets vulnerable on the other.
Unlike previous industrial revolutions that unfolded over decades, the AI economy is scaling at an unprecedented speed.
Khan and Mullin argue that the infrastructure decisions being made now will shape how the industry operates for decades. Once companies and governments build their operations on particular cloud platforms and computing environments, switching becomes expensive and disruptive.
If Canada waits too long to establish its own infrastructure and governance frameworks, the authors warn the country could end up deeply dependent on systems controlled elsewhere.
Much of the infrastructure supporting AI development is concentrated within a small number of global technology companies that operate the cloud platforms and computing environments required to train and deploy advanced systems.
Canada, by contrast, has historically focused on research and early-stage innovation. That strategy helped establish world-leading AI institutes and produced influential researchers whose work helped define the field.
But the report suggests research leadership alone doesn’t guarantee economic leadership.
The authors point to a recurring Canadian pattern where the country generates important scientific breakthroughs but allows the industrial and commercial value to be captured elsewhere.
This disconnect is visible in Canada’s current AI ecosystem. While the country possesses world-class research institutions, promising startups, and expanding compute investments, these elements are not yet working in tandem. There is still no unified industrial strategy to align these pieces for building and operating AI systems at scale.
A significant part of this alignment problem stems from what the report describes as fragmented machinery of government.
Currently, the policies governing Canada’s digital and AI infrastructure are spread across multiple federal departments, which prevents the country from speaking with a single, unified voice.
Khan and Mullin argue that this lack of coordination makes it difficult to negotiate with global technology giants or to manage large-scale infrastructure projects.
To solve this, the authors suggest that Canada needs a more centralized authority or a dedicated digital agency to oversee the “sovereign by design” framework and ensure that investments in compute and cloud services are actually meeting the country’s strategic goals.
Where the AI economy is actually controlled
Much of the public conversation about artificial intelligence focuses on the latest models and research breakthroughs. But Khan and Mullin argue that the real leverage in the AI economy sits deeper in the technology stack.
The systems that train, host, and distribute artificial intelligence are increasingly determining where companies build products, where talent gathers, and where economic value accumulates.
Artificial intelligence systems depend on a stack of technologies that includes computing infrastructure, cloud hosting environments, the models themselves, and the applications that run on top of them.
Each layer concentrates power in different ways.
At the base of the stack is compute, the specialized chips and massive data centres required to train and run advanced AI systems. These facilities require enormous capital investment and long-term operational capacity.
Above that sits cloud infrastructure, which allows companies to access computing power and AI capabilities without owning their own data centres.
On top of the cloud sit the models themselves, the systems trained to generate text, images, predictions, or software code.
And finally come applications, where companies integrate artificial intelligence into products, services, and internal workflows.
Most public attention focuses on the models, but much of the economic power sits in the infrastructure layers underneath them.
Control the infrastructure and companies build their products on your systems. Talent gathers around those platforms. Data flows through those networks. Economic value accumulates within that ecosystem.
The report says the consequences of this are already visible in the global AI economy.
A small group of technology firms now controls much of the infrastructure required to train and deploy advanced AI systems. Their cloud platforms host the tools that thousands of companies rely on to build AI-driven products.
For countries like Canada, that concentration creates a strategic dilemma.
Canada has produced world-class research talent and influential AI institutions. Yet the infrastructure layer where much of the economic value is captured is largely controlled by companies headquartered elsewhere.
The report notes that Canada’s vulnerability is especially bad at the model layer.
While the global market is dominated by a handful of American giants, Canada possesses a unique strategic asset in Cohere, which remains one of the only world-class commercial foundation model companies based outside the United States. Khan and Mullin argue that supporting companies like Cohere is vital for maintaining a sovereign alternative.
Additionally, the authors highlight the growing importance of open-source and open-weight AI models. These systems provide a way for Canadian firms to avoid being locked into the proprietary “black box” ecosystems of foreign providers, which allows for greater transparency and local control over how AI is deployed.
Sovereignty in the AI era isn’t about building a national champion. It’s about making sure domestic firms retain meaningful access to the infrastructure that will shape their industries.
Canada’s strategic choices
Khan and Mullin argue that Canada’s strategy doesn’t hinge on building every layer of the AI stack domestically.
Instead, they point to a handful of areas where policy choices could influence how Canada participates in the global AI economy. One of the most immediate is computing infrastructure.
Training and deploying advanced AI systems requires enormous computing power. Countries that host large-scale compute infrastructure often become hubs for AI research, startup formation, and enterprise adoption.
Canada has already begun investing in this capacity.
The federal government allocated $2.4 billion in the April 2024 federal budget to expand sovereign AI compute infrastructure and ensure Canadian researchers and companies have access to the computing resources required to develop advanced systems.
The leverage: Electricity and minerals
While building domestic capacity is a priority, the report argues that Canada should also use its unique natural advantages to secure its place in the global market. Canada possesses two critical assets that global AI firms desperately need: an abundant supply of clean, low-carbon electricity and a wealth of critical minerals required for hardware production.
Khan and Mullin suggest that Canada can use these resources as strategic leverage. Instead of acting as a passive host for foreign data centres, the country could trade access to its energy grid for guaranteed access to high-end hardware or a commitment to keep certain data and intellectual property within Canadian borders. This would transition Canada from a typical customer to a strategic partner that possesses essential resources.
Beyond direct investment, the report highlights the critical role of government procurement. As one of the largest buyers of technology in the country, the federal government has the power to act as a “first customer” for domestic AI firms.
Currently, much of that spending flows to foreign platforms, which further entrenches Canadian dependence on outside infrastructure. By intentionally directing procurement toward sovereign Canadian providers, the government can help these firms reach the scale they need to compete internationally. This approach would turn public spending into a tool for industrial strategy, ensuring that the tax dollars used to modernize government services also help build Canada’s domestic AI capacity.
Infrastructure is only part of the equation.
The bigger economic impact comes when artificial intelligence is embedded inside major industries such as manufacturing, energy production, financial services, agriculture, and logistics.
Canada’s economy includes globally competitive companies in many of these sectors.
That creates an opportunity for Canadian firms to become leaders in applying artificial intelligence to complex industrial systems rather than competing directly in the global race to build foundational AI models.
The report also emphasizes the importance of international alliances.
Artificial intelligence is inherently global. Few countries can build every element of the technology stack on their own.
Through coordinated investments, shared infrastructure, and aligned governance frameworks, those countries could collectively strengthen their position in the global AI ecosystem and reduce dependence on a small number of dominant technology platforms.
Khan and Mullin argue that by aligning with countries like the United Kingdom, France, and Japan, Canada can move beyond its bilateral relationship with the United States.
These nations face similar structural challenges because they all possess world-class research talent but lack the massive scale of the American tech giants. The authors suggest that a coordinated alliance would allow these middle powers to pool their resources, share computing power, and establish common standards for data privacy and security.
This collective approach gives Canada a seat at the table when global rules for artificial intelligence are being written, which ensures that the country is not simply forced to adopt frameworks designed elsewhere.
Khan and Mullin argue that sovereignty doesn’t mean trying to build every piece of the AI stack domestically. For a country the size of Canada, that wouldn’t be realistic.
Instead, they suggest ensuring Canada maintains enough domestic capability and alternative options that it isn’t locked into a single foreign platform. Supporting sovereign cloud providers, encouraging open-weight models, and maintaining access to multiple infrastructure partners would give Canadian firms room to move if conditions change.
That flexibility matters if prices rise, rules change, or geopolitical pressures reshape how global technology platforms operate.
The report emphasizes that Canada can’t outspend superpowers like the United States or China. While those nations can commit hundreds of billions of dollars to achieve AI dominance, Canada must instead rely on being smarter about how it structures its dependencies.
A central part of this strategy involves preparing for the July 2026 review of the Canada-United States-Mexico Agreement (CUSMA). The authors warn that this review presents a significant risk to digital sovereignty, as Canada may face intense pressure to trade away its ability to regulate its own digital infrastructure in exchange for traditional trade concessions.
To protect its long-term interests, the report argues that Canada must treat its AI policy as a non-negotiable part of its national security during these talks.
What this means for business leaders
For executives, the issue often shows up as a technology decision, but the report argues the implications run much deeper.
Artificial intelligence is rapidly becoming embedded in the core operations of many businesses. Financial institutions use machine learning to detect fraud. Retailers analyze purchasing patterns to refine pricing strategies. Manufacturers rely on predictive models to manage supply chains.
These tools promise efficiency and competitive advantage.
But they also introduce new forms of operational dependency.
When a company runs critical systems on infrastructure it doesn’t control, it becomes subject to the pricing models, policies, and legal frameworks of the platforms providing that infrastructure.
If a Canadian manufacturer, bank, or logistics firm builds core operational capabilities on foreign cloud systems, those systems effectively become part of the country’s economic infrastructure. That also means those operations may ultimately fall under the laws and jurisdiction of foreign governments, something the report notes is already a concern in cases such as the United States’ CLOUD Act.
To understand this risk, the report distinguishes between data residency and data sovereignty.
Data residency simply refers to where the servers are physically located. Many foreign cloud providers have built data centres in Canada to satisfy residency requirements, but Khan and Mullin argue this is not enough.
Data sovereignty, by contrast, refers to which country has the legal authority over that data. Under laws like the U.S. CLOUD Act, the American government can sometimes compel U.S. companies to provide access to data even if it is stored on Canadian soil.
For Canadian businesses and government agencies, this means that physical location does not always guarantee legal protection. True sovereignty requires using providers that are subject only to Canadian jurisdiction.
That raises strategic questions for business leaders:Where are your AI workloads running?
Who governs the platforms processing your data?
What happens if those systems change their pricing, access rules, or regulatory obligations?
These questions go beyond technology strategy. They touch on economic resilience and long-term competitiveness.
The next question is whether the country will help build the systems that run on top of those breakthroughs, or whether Canadian companies will simply plug into platforms developed somewhere else.
Final shots
Canada helped pioneer modern artificial intelligence research. Much of the infrastructure powering the AI economy is now being built elsewhere.
Decisions made in the next several years will determine whether Canada helps shape that infrastructure or primarily relies on systems developed abroad.
The question that is becoming harder to ignore: who controls the systems that increasingly run the modern economy?
Written By Chris Hogg
Chris is an award-winning entrepreneur who has worked in publishing, digital media, broadcasting, advertising, social media & marketing, data and analytics. Chris is a partner in the media company Digital Journal, content marketing and brand storytelling firm Digital Journal Group, and Canada's leading digital transformation and innovation event, the mesh conference. He covers innovation impact where technology intersections with business, media and marketing. Chris is a member of Digital Journal's Insight Forum.
No comments:
Post a Comment