In 1959, Canada cancelled the Avro Arrow, a fighter jet years ahead of anything else in the sky. The program died, the prototypes were cut up for scrap, and a lot of the engineers went looking for work. Many ended up at NASA, helping put Americans on the moon.

Karl Holmqvist, who runs the Canadian defence-tech firm LastWall, brought that up at Arctic Edge, a defence tech event during Toronto Tech Week.

"Canada's really good at building stuff," Holmqvist said, "and we've been really good historically at building even defence stuff."

Building something world-class once is not the same as building an industry that lasts.

The founders on stage had no doubt they could build great products. What they wanted was systemic support to turn those products into an industry. I went into the panel expecting to hear about hardware, and heard about policy for the better part of the morning. The full story is here.

What got my attention was how familiar this all sounded.

Canada was home to a generation of people who built modern artificial intelligence. Geoffrey Hinton did his foundational work in Toronto, Yoshua Bengio in Montreal. The ideas and the talent were ours. 

The companies that turned them into trillion-dollar businesses mostly sit in California, where a research scientist at OpenAI can earn total compensation north of a million dollars a year, according to Levels.fyi, a number that is hard for Canadian labs to match.

When that talent leaves, so does the ownership. As an example, between 2015 and 2020, foreign companies came to own about 70% of the roughly 1,600 AI and machine-learning patents filed by Canadians in the United States, according to the Centre for International Governance Innovation. 

Cohere shows how it's supposed to work. Cohere is headquartered in Toronto, and its co-founders, Aidan Gomez and Nick Frosst, both came out of Hinton's lab. The talent, the company, and the value stayed in the same city. 

The goal, in defence and AI, is to make that the rule instead of the exception.

Some light reading for when you're poking around the new site to see what we changed.

Everyone's using AI, but nobody trusts it One client killed an entire AI project over a rounding error. Jennifer Kervin looks at why trust, not capability, is the thing holding back enterprise AI.

Before you sell your security, close the open door "Hackers don't have to hack in. They log in." Jennifer Friesen wrote about what Arctic Wolf's 2026 numbers say about how breaches actually start.

Toronto Tech Week puts global growth under the microscope Going global doesn't have to mean going south. PhenoTips skipped the U.S. and opted to expand to the U.K. first, and they're doing just fine.

Koho's Interac connection, where cloud-native meets legacy rail Shane Parkhill shared why Koho became the first fintech to build its own direct connection to Interac, and what it saved them.

I'm writing this from San Francisco, where I'm covering Snowflake Summit, the big annual conference for Snowflake, the data and AI company, and I can't help but notice optimism and investment in the city. 

There's an AI-for-everything company in almost every vendor booth. Half the billboards on the way in from the airport are selling AI to people who already build the stuff. 

The whole place runs on the idea that whatever's coming next is worth paying for now.

An ecosystem like this, the energy, the appetite for big expensive bets, isn't something a single government policy can manufacture. Whether the coming AI strategy helps foster one in Canada or gives us another Avro Arrow, we'll have to wait and see. Once it’s officially released, of course. Which has to be soon, right?  

Quick note: this email is late because we launched an updated digitaljournal.com this week, and I didn’t want to send this until the dust settled.

Have a look around the new one and let me know what you think. 

— David

Leadership is asking: are we getting value from AI? Which tools are worth the spend? Where are we exposed? Right now, most teams have no idea.

Harmonic Security Usage Explorer changes that.

You get a complete picture of how your organization uses AI, automatically categorized into custom tasks and use cases.

You’ll see the projects being worked on, who’s using what tools, where AI investments are driving value, and where employees are engaging in risky behavior.

CIOs can rationalize spending and cut wasted licenses. CISOs can pinpoint where risk exists and neutralize it. AI committees can show exactly how their efforts are paying off.

See it in action