Great book to be read by every Canadian entrepreneur and policy maker, even going on 9 years after publishing. IMO the cultural arguments are what are most important here, and they're timeless. It's also a powerful "you can do it" and "we have to do it" book which is motivating.
It was also interesting to note some of the dated arguments in this book which AL is undoubtedly completely aware of. Quick (imperfect) Gemini summary of some of these points:
***
Anthony Lacavera’s How We Can Win (2017) argues that Canada struggles to commercialize its innovations, often selling startups to foreign buyers too early ("branch plant" mentality) and failing to protect intellectual property (IP).
While his central thesis remains relevant, the explosion of Large Language Models (LLMs) and Generative AI (2022–2025) has "dated" specific pillars of his argument in three critical ways:
1. The Shift from "Venture Capital" to "Sovereign Capital"
Lacavera’s 2017 argument focused on the "scale-up gap"—the lack of Series B/C venture capital in Canada to turn $10 million companies into $1 billion companies.
How it is dated: The economics of LLMs have made this gap irrelevant by making it insurmountable for any traditional VC ecosystem. Building a frontier model (like GPT-5 or Claude) now requires tens of billions of dollars in compute infrastructure (GPUs/Data Centers).
The Reality: The game is no longer about "growing a startup"; it is about nation-state level industrial policy. The "winning" strategy Lacavera proposed (better VC incentives) is mathematically insufficient for the AI era, where only hyperscalers (Microsoft, Google) or sovereign funds can afford the entry ticket.
2. The "Data Colony" Problem replaced the "Brain Drain"
Lacavera worried about Canada exporting talent (engineers moving to Silicon Valley) and IP (patents sold to US firms).
How it is dated: LLMs have automated the extraction of value. Canada no longer just loses people; it loses its digital history. US-owned models (OpenAI, Anthropic) have scraped Canadian media, literature, and government data to train their models, effectively colonizing the country's informational heritage without hiring a single Canadian.
The Reality: We are no longer just a "branch plant" where Canadians work for US subsidiaries; we are a "data farm" where Canadian cultural output is harvested for free to improve American products.
3. Obsolescence of Specific Verticals
In How We Can Win, Lacavera placed heavy emphasis on telecommunications (5G) and blockchain as key battlegrounds for Canadian innovation, reflecting his background with Wind Mobile.
How it is dated: While connectivity is important, Generative AI has displaced 5G as the primary driver of economic productivity. The "pipes" (telecom) became commodities, while the "intelligence" (AI) became the value capturer. A strategy focused on "winning" in telecom infrastructure (his 2017 focus) misses the fact that the value has moved up the stack to the application and reasoning layer.
***
Commentary on AI and Sovereignty:
While he hasn't written a dedicated book on AI, his recent speaking engagements and short-form writing have touched on the "Data Colony" concept. He has argued that Canada is repeating its mistakes from the oil and telecom eras by failing to build sovereign AI infrastructure, allowing US giants to once again extract the value.
***
I was curious to see what an LLM would say about what AL's updated pitch would be in late 2025. Would be interesting to know if Anthony agrees:
***
Based on Anthony Lacavera’s recent activities—specifically his July 2025 acquisition of Wealth One Bank—and his established philosophy, his updated pitch to Canadians would shift from "How We Can Win" to "How We Can Survive."
The central argument would evolve from preventing a "Brain Drain" (losing people) to preventing a "Mind Drain" (losing the ability to think independently as a nation).
Here is the updated pitch, merging his 2017 concepts with the reality of the AI era.
The Core Thesis: "Stop Being a Landlord for American IP"
Lacavera would argue that Canada has misunderstood the AI revolution. We are currently positioning ourselves as a "server farm" for the US—using our clean energy and land to power American models (like OpenAI or Anthropic) that we then rent back at a premium. This creates a "Data Colony" where Canada provides the raw resources (energy, data, talent) while the US captures all the value.
Pillar 1: Competition (The "Oligopoly" is Now a Security Risk)
In 2017, he argued that the telecom oligopoly (Bell, Rogers, Telus) hurt consumers.
The 2025 Update: The oligopoly is now a national security threat.
* The Argument: A lack of competition in banking and telecom means Canada lacks the dynamic capital to fund AI infrastructure. The "Big Five" banks and "Big Three" telcos are dividend-dispensing machines, not innovation engines. They are too comfortable to take the risks required to build a sovereign AI ecosystem.
* The Proof: Lacavera is walking this walk by acquiring Wealth One Bank (completed July 2025) to disrupt the banking sector, arguing that we need "Challenger Banks" to fund "Challenger Tech."
Pillar 2: Scaling (Pension Funds Must Come Home)
In 2017, he argued against selling startups too early.
The 2025 Update: He would target the Canadian Pension Plan (CPP) and other major funds (Teachers', OMERS).
* The Argument: Canadian pension funds are among the largest pools of capital in the world, yet they invest heavily in US infrastructure and real estate. Lacavera would argue that this capital must be repatriated to fund Sovereign Compute.
* The Reality: We cannot rely on Venture Capital to build $10 billion data centers. That requires infrastructure money. The updated pitch would demand that Canadian pension dollars build Canadian GPU clusters, ensuring that Canadian startups have access to compute power without paying a "tax" to Microsoft or Amazon.
Pillar 3: Data Sovereignty (The "Epistemic" Crisis)
In 2017, the goal was to keep patents in Canada.
The 2025 Update: The goal is to keep decision-making in Canada.
* The Argument: If Canadian hospitals, banks, and government agencies rely on US-owned black-box models to make decisions, we have lost our sovereignty. We cannot audit the logic of a foreign model.
* The "Truth-First" Policy: Lacavera would likely advocate for a "Made-in-Canada" AI requirement for the public sector—not just "hosted" in Canada, but owned and explainable by Canadians. We need to own the "weights" of the model, not just the building the server sits in.
The Realistic Path: A 3-Step Execution Plan
If Lacavera were to present a roadmap to Ottawa today, it would likely look like this:
* Mandate Open Banking (Immediately):
* Break the data monopoly of the Big Five banks. Allow fintechs and challenger banks (like his own Wealth One) to access consumer data (with permission) to offer better lending rates to entrepreneurs. This unlocks capital for the middle class.
* The "Sovereign Cloud" Tax Credit:
* Instead of subsidizing foreign EV battery plants, redirect that money to subsidize Canadian-owned compute infrastructure. Offer massive tax breaks only if the IP and the hardware are 51% Canadian-owned.
* Weaponize Immigration for AI:
* Wealth One Bank targets newcomers specifically.
Lacavera would argue that Canada’s only advantage over the US is our ability to attract high-skill immigrants. We should offer immediate permanent residency to AI researchers on the condition that they work for Canadian-controlled firms for 3 years, preventing them from using Canada as a waiting room for a US Green Card.
The Bottom Line: The updated pitch is that Canada is currently on track to become a "digital vassal state." The only way to win is to stop renting the future from America and start building the physical and financial infrastructure to own it ourselves.
***
The above are not necessarily AL's position but an attempt to summarize what they might be.
Either way, I recommend the book as a foundational must-read for those wanting to build a better future for Canada.
Happy 2026!