Artificial intelligence may be the most transformative technological development of our generation. It promises to revolutionize medicine, education, manufacturing, scientific discovery and virtually every sector of the economy.
But after months of interviewing artificial intelligence entrepreneurs, economists, investors, cybersecurity experts and technology strategists, I’ve become increasingly concerned that Canada is asking the wrong questions.
We’re fascinated by what artificial intelligence can do. We should be equally concerned about who owns it, who controls it and whether Canadians will share in the economic value it creates.
Because AI isn’t simply a technological revolution. It’s an economic, financial and geopolitical revolution—and potentially a revolution in what it means to be human.
The fundamental question isn’t whether artificial intelligence will change our future. It’s whether Canadians will retain the ability to control that future.
Investors are beginning to question the extraordinary amounts of capital being committed to artificial intelligence.
Technology companies are spending hundreds of billions of dollars developing increasingly sophisticated AI models and constructing the data centres, semiconductor facilities and electricity infrastructure required to support them.
But where are the revenues? Where are the profits? And when will the returns justify these enormous investments?
At the recent Toronto Global Economic Forum, one speaker described equity markets as behaving like teenagers on drugs, while bond markets were behaving like adults.
Another offered an observation that has stayed with me: Markets don’t care until they do. And when they do, they don’t care about anything else.
Economist John Rapley has similarly questioned whether the expected returns from AI will justify the extraordinary spending.
History provides important lessons.
During the nineteenth-century railway boom, investors poured enormous sums into infrastructure that ultimately transformed economies and helped build countries, including Canada. Yet many railway companies failed, leaving investors with enormous losses.
The internet followed a similar trajectory. During the dot-com boom, companies with little revenue and questionable business models commanded extraordinary valuations. When the bubble burst, billions in shareholder wealth disappeared.
Yet the internet transformed the world.
The technology succeeded even though many of the investments failed.
Artificial intelligence may follow a similar path.
We may genuinely need enormous computing capacity, but do we need every data centre being proposed? Will every company developing AI models survive? Will the productivity gains arrive quickly enough to justify the capital expenditures?
As someone who has spent much of his career in corporate finance and real estate development, I’ve learned that an excellent idea can still become a terrible investment if you pay too much, borrow too heavily or overestimate future revenues.
AI is no exception.
During an interview on The Brian Crombie Hour, retired electrical engineer and investor Michael Verge, author of AI Thunder, introduced me to the concept of cascading bottlenecks.
Rather than simply investing in fashionable technology companies, he looks for the essential infrastructure upon which entire industries depend.
In artificial intelligence, those bottlenecks include electricity generation, transmission networks, cooling systems, semiconductors, critical minerals, water and computing infrastructure.
Canada has extraordinary advantages.
Quebec and British Columbia possess substantial hydroelectric resources. Ontario has a sophisticated nuclear industry. Saskatchewan produces uranium. Alberta has abundant natural gas and energy expertise.
We also have internationally respected universities, AI researchers, engineers and relatively stable institutions.
These advantages could position Canada as an important supplier of the infrastructure required for the global AI economy.
But that raises a fundamental question.
Will Canada simply provide the electricity, land and natural resources—or will we own some of the technology and capture a meaningful share of the economic value?
There’s an enormous difference.
Cybersecurity and geopolitical expert Rafal Rohozinski offered one of the most provocative insights during my interviews.
He described artificial intelligence as a geopolitical resource potentially as important to the twenty-first century as oil was to the twentieth.
Once we understand AI as critical infrastructure, the conversation changes.
It’s no longer simply about chatbots, office software or productivity.
It’s about economic power, national security, military capability, information and sovereignty.
Who owns the computing infrastructure? Who controls the data? Who determines how AI models operate? And who can restrict access?
Canada has become deeply dependent on American technology companies for cloud computing, communications, operating systems, cybersecurity and increasingly artificial intelligence.
That dependence has provided enormous economic benefits. American technology companies are among the world’s most innovative and reliable suppliers.
But economic integration and technological dependence are not the same as sovereignty.
Consider satellite communications. Starlink has demonstrated the extraordinary value of satellite internet, particularly during the war in Ukraine. It has also illustrated how much influence the private owner of critical communications infrastructure can exercise.
For Canada, with its vast Arctic territory, remote communities and growing defence requirements, communications resilience is a national security issue.
Similar questions arise with sophisticated military equipment. There is no substantiated evidence that American-built F-35 fighter jets contain a remote kill switch. But modern aircraft depend on foreign-controlled software, maintenance systems, spare parts and technical support.
The relevant question isn’t whether someone can remotely turn off our aircraft. It’s whether Canada can independently sustain critical military capabilities if access to foreign support becomes restricted.
Increasingly, the same concerns apply to civilian technology.
During my interview with Belgian technology entrepreneur Jo Caudron, we discussed Europe’s growing interest in digital sovereignty and alternatives to excessive dependence on American cloud services and software.
Canada doesn’t need to develop domestic alternatives to every foreign technology.
But we should have credible alternatives for essential government systems, sensitive data, communications, financial infrastructure and national security.
Sovereignty doesn’t mean isolation. It means having choices.
At the recent Urban Land Institute conference in Miami, I attended a panel examining the economics and challenges of data centre development.
Initially, participants identified electricity availability and zoning as the principal obstacles.
By the end of the discussion, however, community support had emerged as a central concern.
Communities are asking legitimate questions.
Will data centres increase electricity costs? How much water will they consume? How many permanent jobs will they create? What tax revenues will they generate? And what benefits will local residents receive?
The panel highlighted Quincy, Washington, where data centre investment has helped support public infrastructure and community services.
These benefits matter. Data centres can create substantial construction employment, skilled technical jobs and municipal revenues.
But the benefits must be demonstrated rather than assumed.
And there is another issue Canadians should understand.
A foreign technology company can build a massive data centre on Canadian land, use Canadian electricity, employ Canadian workers and pay Canadian taxes while retaining complete control over the computing capacity, software and commercial decisions.
Such a project may still be economically valuable.
But it doesn’t necessarily strengthen Canada’s technological independence.
Geography is not the same thing as sovereignty.
Before governments subsidize or approve major data centre developments, they should understand both the community benefits and the implications for Canadian control over critical digital infrastructure.
Jo Caudron raised another important possibility.
Rather than trying to outspend the United States and China developing increasingly powerful AI models, perhaps countries like Canada should focus on becoming among the world’s most effective users of artificial intelligence.
Consider the possibilities.
AI could help reduce healthcare waiting times, improve manufacturing productivity, accelerate scientific research, develop mineral resources more efficiently, modernize agriculture and improve government services.
These applications could generate enormous economic and social benefits.
Perhaps the greatest beneficiaries of the AI revolution won’t be the countries that build the smartest models, but those that use artificial intelligence most effectively.
Canada should pursue both objectives: become an exceptional adopter of AI while maintaining sufficient independent computing capacity and technological expertise to protect our sovereignty.
The challenges aren’t exclusively economic or geopolitical.
Psychologist Dr. Gleb Tsipursky, author of The Psychology of AI Adoption at Work, explained during our interview that many barriers to successful AI adoption are psychological rather than technological.
Employees fear displacement. Organizations fail to provide adequate training. Poorly implemented systems can increase workloads rather than reduce them.
Investor John Ruffolo has raised concerns about how quickly AI could disrupt employment across professional occupations, including law, finance, accounting, programming and research.
What happens if technological change outpaces our ability to retrain workers? What happens if productivity increases but wages and employment opportunities fail to keep pace?
We could create a wealthier economy while leaving millions of people feeling less economically secure.
That would represent a failure of public policy.
And there is another, more personal concern.
During my interview with Dr. Thomas Verny, we discussed AI companions capable of conversation, emotional reassurance and simulated empathy.
Such systems may provide genuine comfort to lonely or isolated individuals.
But can simulated companionship replace human relationships?
I’ve spent years studying social capital—the connections, conversations, compassion, collaboration and communities that create trust and belonging.
I call it the Power of Co.
I believe these relationships will become more important, not less important, as artificial intelligence becomes increasingly sophisticated.
Intelligence isn’t wisdom. Information isn’t understanding. And simulated companionship isn’t love.
Canada was built through extraordinary infrastructure investments.
The transcontinental railway. The St. Lawrence Seaway. Our highways, hydroelectric systems and nuclear industry.
These projects connected our communities, strengthened our economy and increased our independence.
Perhaps the nation-building projects of our generation will be digital.
Canadian computing capacity. Secure communications networks. Reliable electricity infrastructure. Canadian-controlled data. AI research and commercialization.
But we need a national strategy guided by five principles: productivity, sovereignty, infrastructure, people and international partnerships.
We should invest in technologies that measurably improve economic performance. Protect critical digital systems. Build infrastructure that creates lasting value. Help workers adapt. And strengthen technological partnerships with Europe, Japan, Australia, South Korea and other trusted countries.
We must also maintain financial discipline.
Not every data centre deserves public subsidies. Not every AI investment will generate adequate returns. And strategic importance is not an excuse for financial irresponsibility.
Canada needs to distinguish between participating in a technological revolution and blindly financing an investment boom.
Ultimately, this is about much more than artificial intelligence.
It’s about what kind of country we want to become.
Will Canada export its talent, energy and natural resources while importing the technology that controls its future?
Or will we use those advantages to build greater economic strength and technological independence?
Will we become increasingly dependent on decisions made in Washington, Silicon Valley and Beijing?
Or will we develop the capabilities and partnerships necessary to make our own choices?
And will AI strengthen our communities and human relationships—or leave us more technologically sophisticated but increasingly isolated?
We don’t build countries simply with computers, capital, electricity or artificial intelligence.
We build countries with people. With trust, cooperation, shared purpose and the willingness to work together.
That’s how we built Canada before.
And that’s how we need to build it again.
We don’t need to own every technology in the world.
But we do need to own our decisions.
Because the greatest danger isn’t simply that artificial intelligence becomes more powerful than we imagined.
It’s that we become more dependent than we intended—and discover too late that we’ve surrendered the ability to choose our own future.
Brian Crombie is a corporate finance and real estate executive and host of The Brian Crombie Hour on Sauga 960 AM, where he interviews leaders in business, economics, technology, politics and public policy.
