The Canadian hardware playbook isn't the software playbook

In January, Montréal's Vention closed a $150-million round. The detail that matters isn't the size. It's who wrote the cheques: Investissement Québec put in $55 million, its largest venture bet to date, and more than 75% of the round came from Canadian sources. The CEO said the company wanted to make sure Vention "stays Canadian forever."

A month later, another Montréal robotics company, Mecademic, raised $21 million in a round led by Investissement Québec, with Export Development Canada and BDC participating. In May, a Waterloo quantum hardware spin-off called QuantumCore announced $10.7 million, including $1.7 million from an NSERC Alliance grant that gives it access to a university lab.

Put those side by side with the software story we wrote about Vancouver's AI talent gap, where the company tends to follow its US lead investor south. Canadian hardware teams seem to be running a different playbook. It isn't better in every way, but software founders and PMs can learn a lot from it.

The TL;DR:

• The best Canadian hardware teams build a Canadian capital stack: grants, university facilities, then crown and provincial investors, before (or alongside) private VC.
• They sell to technical buyers early (factories, labs, other hardware companies), not to consumers.
• Because their money, labs and assembly are here, they tend to stay here. That's the opposite of the software pattern.
• 2026 added a demand side: Buy Canadian procurement rules and a defence strategy aiming for 70% domestic contracts.
• The weak spots are real: the scale-up funding gap, overseas supply chains and competing with California for talent.

Hardware founders treat government money as a stage, not a side quest

A typical Canadian software startup thinks of SR&ED and IRAP as nice-to-haves that extend runway. The hardware teams in our examples treat public money as a deliberate stage of the company.

Look at the order of operations. QuantumCore spun out of the University of Waterloo's Institute for Quantum Computing. It uses its co-founder's lab and the university's nanofabrication facility, which no seed round could buy as cheaply. Its CTO put it this way: "Canada has this homegrown way of financing ventures like quantum tech."

Then comes patient, state-linked capital. Investissement Québec, EDC and BDC showed up in both big Montréal rounds this year. They're investors that expect a long build and care about where the jobs land. For a company that needs tooling, certification and inventory, that patience matters more than a famous logo on the cap table.

Haply Robotics, another Montréal company, was candid about this at CES last year: its co-founder credited SR&ED with easing the company's work. That's a hardware founder talking about a tax credit as a core part of the business, not a footnote.

Why does this work better for hardware than software? Our read: hardware R&D is expensive, slow and easy to document as R&D. Those are precisely the costs these programs are built to cover. A software team iterating weekly on a SaaS product gets less out of the same programs.

They sell to people who buy tools, not apps

The second pattern is who the first customer is.

Vention sells to engineers. Its customers are engineers at more than 4,000 factories in 30 countries, including Toyota, GE, Tesla, Siemens and Airbus, and it has passed $100 million in annualized revenue. Mecademic makes compact, high-precision industrial robots. Qubic, a Sherbrooke company spun out of Sherbrooke's Institut quantique and Waterloo's IQC, signed its first amplifier customer, US-based Quantum Machines, in May, and its CEO points to demand from quantum computing and defence buyers.

That's business-to-business, often business-to-engineer. It suits Canada for a simple reason: a small home market hurts much less when your buyer is a factory in Germany or a lab in Boston, who will evaluate you on specs, not brand.

Notice too that QuantumCore's CEO said they aren't trying to build the most powerful quantum computer. Both quantum companies sell the components everyone else needs. That's a product choice worth stealing: be the pick-and-shovel supplier in a race, not one of the racers.

Because the money and the factory are here, the company stays here

This is the part that should interest everyone, not just hardware people.

In our Vancouver piece we cited a Leaders Fund study, reported by The Globe and Mail, finding that just 32.4% of Canadian-led high-potential startups launched in 2024 were headquartered in Canada. Software companies move to where the lead investor and the early customers are.

Hardware companies have heavier anchors. Haply does large portions of production, including final assembly, in Canada. Montréal's Eli Health says it does all its research and manufacturing here. When your lab, your assembly line and your largest investors are local, moving the headquarters to Delaware and the team to San Francisco is a much bigger decision than changing a mailing address.

We don't have a clean dataset comparing headquarters moves for Canadian hardware versus software startups, so treat this as a pattern in the examples, not a measured fact. But it's a hypothesis worth testing: hardware may be the part of Canadian tech where the headquarters leak is smallest.

Waterloo and Montréal get there differently

The two cities aren't running identical plays.

Waterloo leans on the university. Its strengths are spin-outs from research institutes like IQC, co-op talent that has shipped real engineering work before graduating, and incubator space built for physical products. Velocity, the university's incubator, has had a dedicated hardware, materials and life sciences space, the Foundry, since 2014. The Waterloo playbook is technical depth first: grants and lab access, then capital.

Montréal leans on the province. Quebec runs an active industrial policy through Investissement Québec, which has invested $800 million in venture capital over the past decade, plus Desjardins and the crown corporations. The Montréal playbook is industrial scale with a patient partner: robotics and automation companies that sell into manufacturing, backed by investors who want the headquarters to stay.

And the two talk to each other. Qubic is literally a joint spin-off of a Quebec and a Waterloo institute.

2026 handed hardware teams a customer: the government

Until this year, the public money was mostly on the supply side: grants and investment. Now there's a demand side.

The federal Buy Canadian Policy took effect December 16, 2025. Canadian suppliers get extra points in bid evaluation, and more points for Canadian content such as domestic manufacturing and R&D. It applies to contracts of $25 million and up, dropping to $5 million, and a Small and Medium Business Procurement Program was promised for spring 2026.

Then in February, the federal defence industrial strategy set a goal of awarding 70% of defence acquisition contracts to domestic companies, up from about 43%, with research money aimed at AI, robotics, autonomous systems and quantum.

Read those lists again. Robotics and quantum are exactly where the Waterloo and Montréal examples above sit. Canadian content points reward exactly what hardware teams already do: build and assemble here.

Two cautions. First, procurement is slow, and a startup that bets its runway on a government contract can die waiting. Second, the industry has been asking for this for a while. A ventureLAB roundtable last year argued that government buying could validate early-stage firms and that revenue is the best growth funding. We don't know yet whether the new rules will actually reach small companies, or mostly benefit large primes and foreign-owned subsidiaries with Canadian plants.

Where the playbook breaks

It's not all working.

The scale-up gap. The same ventureLAB roundtable said investment in hardware companies was already declining, and that nondilutive support largely disappears once a company starts raising and earning revenue. The Canadian stack is great from zero to prototype. It's thinner from prototype to volume.

Supply chains. Not everyone can build here. The founder of home-composter maker VCycene told BetaKit at CES that it manufactures in China for fast access to a supply chain that isn't available in Canada, and that component suppliers tend to stay overseas even when factories are domestic. With tariffs shifting, that's a product risk, not just an operations one.

Talent. Waterloo micro-semiconductor startup VueReal said at the same event that it's hard to attract people who would otherwise go to California.

What you can do tomorrow

PMs (including software PMs)

• Learn the capital stack. Ask your finance lead how much of your R&D is claimed under SR&ED and what would change if the roadmap were documented with that in mind. Product decisions affect what's eligible.
• If you work on hardware, treat the bill of materials and the supplier list as part of the product spec. A part that ships from one overseas supplier is a roadmap risk.
• Try selling to engineers. If your product can have a technical buyer who judges on specs, a small home market matters much less.

Founders

• Sequence your money on purpose: grants and university facilities first, then patient capital (provincial funds, BDC, EDC), then private VC for scale. Don't raise dilutive money for work a grant or a lab partnership would cover.
• Read the Buy Canadian rules now, and work out what your Canadian content score would look like. Just don't build a runway plan that depends on a government purchase order arriving on time.
• Consider the pick-and-shovel position. Both quantum startups above sell components into a race rather than entering it.

Engineers

• Hardware and robotics teams in Waterloo and Montréal are hiring for the defence and automation wave. Firmware, controls, test and manufacturing engineering are skills that transfer and that software-only folks rarely have.
• If you're at a university, find out what lab and fab facilities startups can use through grants like NSERC Alliance. That's a cheaper first year than any seed round.

The software playbook says raise in the US, sell in the US, and the company follows. The hardware playbook these teams are running says build the stack here, sell to whoever needs the part, and stay. Neither works for everyone. But if Canada wants more companies that stay, it should study the ones that already do.