Canadian AI: Defence AI has a moment, and 93% of companies are using AI but only 2% can prove it's working.
And more on $79.5M for AI manufacturing, Shopify's AI in engineering playbook, and five AI building reservists who impressed NATO.
Good morning! Welcome to the Canadian AI Newsletter, a weekly rundown for founders, operators and investors.
Was this forwarded to you? Subscribe below.
I am Raif Barbaros, Partner at Mistral Venture Partners. Views are my own.
This was the week defence AI went from “emerging theme” to “dominant storyline.” Lockheed Martin wrote a $3.6M cheque to an Ottawa AI shop. Five Canadian reservists built an LLM-powered cyber tool that was the only AI capability at a NATO exercise. Ottawa pushed $13.8M into BC defence AI startups. And Clearpath Robotics’ co-founder sat down with BetaKit to explain why Canada needs more robots, not fewer, in its military. Meanwhile, the federal government put $79.5M behind AI in manufacturing, Shopify revealed the internal engineering playbook that explains how it keeps shipping without adding headcount, and KPMG dropped a stat that should make every enterprise AI vendor uncomfortable: 93% of Canadian organizations are using AI, but only 2% can show a return on it. The money is flowing. The question is whether the results will follow. Let’s get into it.
💰 Deals & Milestones
Lemay.ai (Ottawa) secured a $3.6M investment from Lockheed Martin Canada for collaborative AI R&D in defence and aerospace.
Focus areas: predictive aircraft maintenance, supply chain optimization, GPS-denied navigation, and “sovereign knowledge management” for the Canadian Armed Forces.
Investment enabled by Canada’s Industrial and Technological Benefits (ITB) policy, tied to the CC-130J Super Hercules fleet. Includes collaboration with the Czech Aerospace Research Centre.
Lemay.ai is a small Ottawa shop (founded 2015 by CEO Matt Lemay) that has delivered AI solutions to NATO and defence clients. Minister Joly framed this as “how defence investments can strengthen national security while driving economic growth at home.”
Defence AI keeps landing in Ottawa. Between this, Denvr’s two partnerships (Issue 7), Larus Technologies’ $8.3M contract (Issue 7), and Dominion Dynamics’ $50M commitment (Issue 4), the capital’s defence-AI cluster is becoming hard to ignore.
WELL Health Technologies (Vancouver, TSX: WELL) partnered with AliveCor to bring AI-powered cardiac monitoring to Canadians.
AliveCor’s Kardia platform uses Health Canada-licensed AI algorithms to detect three common heart arrhythmias from 30-second ECG recordings. WELL’s Canadian-registered cardiologists provide clinician reviews within 24 hours.
Addressing a real bottleneck: elective cardiology wait times are up 53%, with Canadians waiting an average 15.3 weeks for specialist consultations.
Future phases may deploy AliveCor’s pocket 12-lead ECG across WELL’s 250+ clinic network. AI tackling healthcare wait times is exactly the kind of real-world application that moves the needle.
Aspect Biosystems (Vancouver) announced a $280M partnership with the Government of Canada, including $79M in federal investment through the Strategic Response Fund.
UBC spinout building AI-powered 3D-bioprinted tissue therapeutics, primarily targeting Type 1 diabetes. Partnership with Novo Nordisk (Ozempic maker) since 2023.
~130 employees, $250M+ USD raised to date. One of the largest single federal investments in a Canadian biotech company.
The AI component powers the bioprinting platform itself. The $79M is a significant bet on Vancouver’s deep-tech corridor.
AXL (Toronto) appointed nine U of T professors as its inaugural Faculty Fellows cohort, bringing industry expertise from Nvidia, NASA, Samsung, Adobe, Intel, and Microsoft.
Notable fellows include Gennady Pekhimenko (Nvidia senior director of AI software, CentML co-founder, Vector Institute faculty), Sven Dickinson (former head of Samsung Toronto AI Research Centre), and Steve Easterbrook (former NASA lead scientist).
Fellows compensated with “sweat equity” units across every studio company. AXL has approved nine investments to date and is on pace for its goal of launching 50 AI companies in five years.
CEO Daniel Wigdor is explicitly framing this as a brain drain countermeasure. The calibre of this cohort says something about the pull of Toronto’s AI ecosystem when the incentive structure is right.
ThinkLabs (New York / Toronto) closed a $28M USD Series A to help utility providers modernize power grids under strain from AI data centre demand.
Led by Energy Impact Partners, with NVentures (Nvidia’s VC arm) and Edison International. Previously raised $6.8M CAD seed from GE Vernova.
Uses “physics-informed AI” to build digital twins of power grids. Compresses month-long infrastructure studies into under three minutes. Runs 10 million scenarios in 10 minutes with >99.7% accuracy.
Founded by UWaterloo Engineering alum Josh Wong, who previously led Toronto Hydro’s smart-grid division and founded Toronto-based Opus One Solutions (sold to GE in 2022). ThinkLabs maintains a Toronto office and is actively hiring in Canada.
🏢 Large Company / Big Tech
Shopify (Ottawa/Toronto) revealed its AI-first engineering playbook in a detailed Bessemer Venture Partners interview with VP & Head of Engineering Farhan Thawar.
Key detail: Shopify built a centralized LLM proxy routing all AI requests through a single platform layer. Engineers now run multiple AI agents in parallel on different parts of the codebase, then review and merge outputs.
Engineering team estimates ~20% productivity gains from AI. Thawar described the 2026 shift as “agentic harnesses” and warned: “If you don’t figure out how to harness agents in 2026, you’ll be behind.”
Shopify has now appeared in all eight issues of this newsletter. At this point I should just give them a recurring column. But this one is different from the merchant-facing product announcements we’ve covered before. This is about how Shopify is rewiring its own engineering DNA. The “no net new hires while growing revenue” story (Issue 4) now has a technical explanation.
🔬 Research
CAFCYBERCOM (Canadian Armed Forces Cyber Command) showcased a homegrown AI-enabled cyber tool at NATO’s Coalition Warrior Interoperability Exercise (CWIX) in Poland.
A five-member team from 33 Signals Regiment built an LLM-powered Cyber Indications and Warning tool for cyber analysis, translation, and geolocation. Canada was the only participant in the cyber portion to employ an AI-driven capability. Selected for CWIX Innovation Spotlight.
Built by reservists working Class A time since January 2025. The tool aims to integrate cyber intelligence into the National Common Operating Picture.
LCol Liam Robertson: “This project shows what is possible when you empower small, motivated teams to experiment and deliver.”
Five reservists. Class A time. An LLM tool that was the only AI capability at a NATO exercise. This is the kind of story that makes you proud of what scrappy Canadian teams can build with minimal resources.
🏛️ Policy
NGen (Next Generation Manufacturing Canada) announced $79.5M for 20 AI projects to help Canadian manufacturers adopt AI. Announced at the N³ Summit in Toronto, with AI Minister Evan Solomon in attendance.
$29.2M in new federal funding through the Pan-Canadian AI Strategy, combined with $50.3M from industry partners. The ~2:1 private-to-public co-investment ratio shows strong industry buy-in.
Named participants include e-Zinc + Katalyze AI (AI battery quality control), Xaba + Martinrea International (AI vision robotics for powerpacks), InPho/ElectroPhotonic-IC (AI in semiconductor manufacturing), and Magna International (applied AI robotics).
Canada’s manufacturing robotics adoption rate sits at only 8.4%, below Thailand and Mexico. NGen CEO Jayson Myers: “These projects are about turning Canadian AI into Canadian productivity.”
This is the Pan-Canadian AI Strategy shifting from research to production. The fact that Magna is in the mix signals that serious industrial players see AI manufacturing as a near-term operational upgrade, not a science project.
PacifiCan invested $13.8M in five BC projects through the Regional Defence Investment Initiative (RDII), advancing AI and aerospace technologies.
AI-specific investments include: $1.4M to UVic’s Advanced Control and Intelligent Systems Lab for an AI-powered autonomous drone mapping system; $2.4M+ to Atreides for an AI-enabled unmanned systems data platform; $2.8M+ to OSI Maritime Systems for AI-augmented collision avoidance software.
Part of the $379.2M national RDII under the Defence Industrial Strategy. Three of five funded projects have explicit AI applications.
Defence AI funding is no longer concentrated in Ottawa. Vancouver Island is now on the map. Atreides and OSI Maritime Systems are names worth watching.
Michael Schull (President & CEO, Digital Research Alliance) and Feridun Hamdullahpur (former UWaterloo president) authored a Hill Times op-ed arguing that sovereign Canadian AI requires investments beyond compute infrastructure.
Key argument: “Sovereign AI requires more than infrastructure; it depends on data and talent to translate capacity into real-world impact.”
The critique complements last week’s budget euphoria. The money is flowing, but without the data ecosystems and talent pipelines to back it up, sovereign compute becomes expensive empty infrastructure.
📊 Data
KPMG Canada published “Beyond AI Adoption: Turning Canada’s AI Momentum into Measurable Returns”, surveying Canadian business leaders.
93% of organizations now using or piloting AI, up from 61% the previous year. But only 2% report measurable ROI. Only 31% have fully integrated AI across core operations.
Canada ranks 42nd of 47 countries in AI trust and 44th of 47 in AI literacy. Only 29% of employees say their employer has a comprehensive AI use policy.
The 93%/2% gap is the stat of the week. Nearly everyone is experimenting. Almost nobody can prove it’s working. Last issue, MNP found 91% satisfied with AI progress but only 4% called it “transformational.” The pattern is clear: adoption is outrunning outcomes. The next wave of enterprise AI value gets created when that gap closes.
CFIB (Canadian Federation of Independent Business) reported that 45% of Canadian small businesses are now using generative AI.
Usage scales with firm size: 39% for firms under 5 employees, 60%+ for firms with 20-49 employees.
78% of firms plan to maintain or increase AI training spending in 2026. AI investment and employee training spending are moving in lockstep.
Small business AI adoption doesn’t get enough coverage. The 45% figure for SMBs is remarkably high and suggests the AI wave is reaching well beyond the enterprise tier.
In brief
Globe and Mail updated its newsroom AI policy: staff and contributors are now prohibited from using AI to edit or write any part of a story. The piece also surfaced the “Victoria Goldiee” saga, where an AI-generated freelancer was submitting work to a Canadian magazine. Canadian journalism is drawing hard lines.
Natural Resources Canada opened applications for its “AI for Canadian Energy Innovation” program. Grants of $500K to $1.5M per project, covering up to 75% of costs, running through March 2030. Fresh federal money for energy-AI R&D.
ConstructConnect published a major feature on Canada’s AI data centre construction boom. Key numbers: Bell’s $1.7B Saskatchewan facility (300 MW, Canada’s largest), eStruxture’s $750M CAL-3 in Rocky View County, Alberta, and proposed Alberta data centre load of 21 GW against a provincial peak grid capacity of 12.8 GW. The physical layer of Canadian AI keeps getting bigger.
Canadian Mining Journal explored how AI data centre demand is rewriting mineral markets for Canadian miners. Copper, lithium, rare earths. Canada’s geology, governance, and proximity to the US position it well in the AI supply chain. Read about how AI is driving this demand all the way down to what’s underground.
Clearpath Robotics co-founder Ryan Gariepy sat down with BetaKit for a Q&A on lethal autonomous weapons and Canada’s defence strategy. Gariepy, who led Clearpath from founding through its ~$600M USD acquisition by Rockwell Automation, now chairs the Canadian Robotics Council. His position: he supports military robots for logistics, recon, and even weaponized applications with proper controls, but draws the line at fully autonomous lethal decision-making. On Canada’s opportunity: “We have a lot more space, a lot fewer people, and our environment is a lot more hostile. That is the perfect place for robotics.” Worth the full read.
📅 Upcoming Events
ALL IN Vancouver — April 15, 2026
SASK AI EXPO (Saskatoon) — April 27, 2026
The AIA (Montreal) — May 5, 2026
Upper Bound (Edmonton)— May 19–22, 2026
ALL IN Toronto — May 28, 2026
Canadian AI Conference 2026 (Vancouver) — May 25-29, 2026
ALL IN Main Event (Montreal) — September 16-17, 2026
Thank you for reading! Subscribe, and please share any feedback 🙏
— Raif Barbaros