More Innovation Funding, More Orphans: Why Canada's Valley of Death Is Widening

In its 2026 study of how Canada buys defence equipment, the Office of the Procurement Ombud records a line from the experts it interviewed that should stop any founder holding an innovation award: "The 'valley of death' is expanding: As Canada increases the number of innovation funding programs, many more 'orphan' products and services will fail to reach procurement" (Office of the Procurement Ombud, 2026).

Most people will read Canada's new defence money as the bridge across the valley. The experts the Ombud interviewed read it as more traffic into it.

Government money for the industry's own research is real and growing. The share of defence-industry R&D funded by Canadian government sources rose from close to 15% in 2022 to nearly 25% in 2024 (Innovation, Science and Economic Development Canada [ISED], 2026). What research money buys is a prototype. Whether anyone then buys the product is a separate decision, made by different people, on a different clock.

Funding a prototype is not buying it

The defence procurement experts the Ombud interviewed name three causes. Despite investment in early-stage innovation through the IDEaS program, "many promising technologies fail to transition into CAF programs due to a lack of sustained funding, unclear adoption pathways to procurement and lengthy procurement timelines" (Office of the Procurement Ombud, 2026).

The third cause is the one a small firm feels first. The same report records that "SMEs often cannot absorb the financial strain associated with the long timelines required for Canada to procure products emerging from government-funded R&D" (Office of the Procurement Ombud, 2026). A grant ends on a date. A procurement starts when a buyer is ready. The months between the two are paid for by the founder.

None of this is unique to Canada. Jeff Decker and Eric Li, writing about software firms entering the U.S. defence market, describe companies that "mistake non-recurring revenue from the Pentagon as success and then get trapped in the 'valley of death'" (Decker & Li, 2023). They treat the one-time money as training wheels: a way to find defence product-market fit and build relationships before going after recurring revenue. My addition is timing — given how long a Canadian procurement takes, I would start building that route while the first award is still running, not after it ends.

Crossing takes three people, and a prototype usually has one

The best evidence I know on what gets a project across comes from inside companies, not governments. Stephen Markham and colleagues surveyed 272 experienced members of the Product Development & Management Association about the front end of innovation. They found that three roles move a project across the valley together: champions "make the organization aware of opportunities by conceptualizing the idea and preparing business cases"; sponsors "support the development of promising ideas by providing resources to demonstrate the project's viability"; and gatekeepers "set criteria and make acceptance decisions" (Markham et al., 2010).

Their study was of firms. My reading is that the same three roles decide a sale into government, and that a funded prototype usually has only the first Sagentix GTM Methodology, 2026. The founder is the champion. The sponsor is whoever holds a budget line and will spend political capital on your product. The gatekeeper is the contracting authority and the security regime behind it.

In Canada's security regime the word is literal. "An eligible organization must be sponsored to get security screened" (PSPC, 2026c). On subcontracting, the same page adds that "Organizations cannot request a pre-clearance under any circumstance." Approved sponsors include federal procurement officers and cleared companies that already hold a contract with security requirements; the page also describes a route for under-represented organizations to be sponsored before a procurement. In the ordinary case, though, a sponsor has to exist before the gate will open.

The market side: name who pays, then name the route

Three things get called a win that are not the same thing: a qualification, which gives you the right to compete; funding, which pays you to advance a technology for a program office; and a contract, which comes with an invoice. Only the third is revenue Sagentix GTM Methodology, 2026. Sort every item in your pipeline into one of the three before you forecast anything.

Then look at where the revenue in this industry actually sits.

Canadian defence industry, 2024: share of firms vs share of revenue

• Under 250 employees — share of firms86%
• Under 250 employees — share of revenue20%
• 500+ employees — share of firms8%
• 500+ employees — share of revenue70%

Firms under 250 employees are 86% of the industry and earn 20% of its revenue; firms with 500 or more employees are 8% of firms and earn 70% (ISED, 2026). The Ombud draws the conclusion from the earlier edition of the same data: small firms "primarily participate in the defence sector as suppliers to larger companies in supply chains" (Office of the Procurement Ombud, 2026). Decker and Li point the same way for the U.S. market, where they name the prime contractors as the most common channel partners (Decker & Li, 2023).

So the first buyer to name is often not a department. It is the supplier-development or engineering lead at the prime whose program needs what you built. That person is also, in the Markham sense, the likeliest sponsor you can find this quarter.

The gate side: start the steps you can take alone

Some gates wait for a sponsor. Two do not, and both are cheaper to start early than to rush later.

Controlled Goods registration. The Controlled Goods Program requires registration before you "receive bid solicitation documents containing controlled goods or controlled technology" (PSPC, 2026a). There is no fee: "There is no cost to register in the CGP," though the page adds that compliance with the Defence Production Act can carry administrative costs. And the queue is long right now; the program says its processing times "are now significantly longer than our 32-business day standard" (PSPC, 2026a). A firm that waits for a solicitation before applying can be locked out of the documents it needs to bid.

CPCSC Level 1. Where a defence contract calls for Canada's new cyber certification, it binds later than most people expect: "Level 1 self-assessment will be required at contract award, and not during the bidding process" (PSPC, 2026b). That is the moment you have the least time and the most to lose. I covered the mechanics in an earlier piece on the award gate.

Dual-use is the best way across, and it makes two claims

Dual-use technology has both civilian and military uses, and Canada's new money is aimed straight at it. The Defence Industrial Strategy, launched on February 17, 2026, commits an initial $656.9 million "to support the development and commercialisation of defence and dual-use technologies" and a new $4 billion Defence Platform at the Business Development Bank of Canada (Prime Minister of Canada, 2026). In January, NRC IRAP announced Defence Industry Assist — $241 million in advice and funding for SMEs "to advance made-in-Canada defence and dual-use technologies" (National Research Council Canada, 2026).

My verdict: dual-use is the strongest route across the market side of the valley, because a commercial customer can pay the bills while a defence buyer decides. But a founder who says "we're dual-use" is making two claims — and each needs its own evidence.

The market claim: someone outside defence pays you. A first award does not make that happen by itself. Carlo Bottai, Gaétan de Rassenfosse and Emilio Raiteri traced 3,070 patents from the U.S. Department of Defense's Small Business Innovation Research (SBIR) program, looking on the web for signs that each one reached a product. They found signs for 21.5%, against 18.5% for a benchmark group of comparable patents (Bottai et al., 2025). For projects that stopped at Phase I, the first and smaller award, "the difference between the SBIR-funded and the benchmark group is never statistically significantly different from zero"; the lift came from projects that went on to Phase II (Bottai et al., 2025). That is U.S. data. A Canadian study of IDEaS or Innovative Solutions Canada showing otherwise would change my mind; I have not found one.

Jordi Molas-Gallart's older work explains why the second step is hard. Writing about defence firms in the 1990s, he found wide agreement that dual-use policies mattered, "but no agreement seems apparent on what they are" (Molas-Gallart, 1997). He separated "straight" transfers, where a technology moves to a new use unchanged, from "adaptational" ones, where it has to be reworked, and called the second kind "more ambitious policies and riskier" (Molas-Gallart, 1997). For a founder, my reading is that the commercial version of a defence prototype is usually a second product (with its own buyer, price and support), and the plan should say so.

Investors price that option. Among listed European and U.S. defence contractors, Marco Albori, Fabrizio Ferriani and Livia Ristuccia found that the stock-market premium on R&D "rises with dual-use (non-strictly military) exposure" (Albori et al., 2026). Those are large public companies, so the finding says nothing directly about a startup's next round. It does show the market paying more for research that can sell on both sides.

The legal claim: which list your product is on. In export law, "dual-use" is a list, not a label. Group 1 of Canada's Export Control List is the Dual-Use List. Its categories run from electronics and computers to information security (encryption), sensors and navigation, and unless the guide says otherwise its controls "apply to all destinations except the United States" (Global Affairs Canada, 2026). So a commercial sale to a European customer can need an export permit — with no military buyer anywhere in the deal. The reverse also surprises people. The Controlled Goods Program covers items in Group 2 (the munitions list), item 5504 of Group 5, Group 6 and U.S. ITAR defense articles, not the Group 1 Dual-Use List (PSPC, 2025). A purely dual-use product can need an export permit and no Controlled Goods registration — until a defence bid puts controlled technical data in your hands (PSPC, 2026a).

Where I fit, and what I do not do

I work both sides of this valley. On the market side, a Sagentix engagement names the buyer, the sponsor and the route, sourced and checked through an 18-check quality gate, in engagements priced from CA$4,500 to CA$45,000; the CA$4,500 Phase 1 carries a money-back guarantee, subject to terms. On the gate side, Sagentix prepares defence and dual-use suppliers for the assessments and registrations that stand between a pilot and a contract — Controlled Goods registration, CPCSC readiness, Protected B authorization and AI governance. That work is scoped and quoted separately.

Sagentix Advisors prepares organizations for independent assessment and authorization. Sagentix is not a 3PAO: we do not perform the independent assessment, issue certification, run penetration tests, or operate a SOC. Where independence is mandated, that role belongs to your assessor. We author the evidence; an independent party judges it.

Three ways to act on this

1. Do it yourself. Sort your pipeline into qualification, funding and contract. Name one sponsor at a prime for each product. Classify your product against the Export Control List. Start the Controlled Goods application if your bids will touch controlled technology. None of this needs an outside advisor.
2. Go straight to the source. The Controlled Goods Program and the Contract Security Program both publish their requirements and contacts on Canada.ca, and Controlled Goods registration is free (PSPC, 2026a). Global Affairs Canada publishes the Export Control List guide (Global Affairs Canada, 2026). Read the Ombud's study in full; it is the clearest public account of why good technology stalls.
3. Work with Sagentix. Start with the CA$4,500 Phase 1 for the buyer map, money-back subject to terms; the gate work is scoped separately.

References

• Albori, M., Ferriani, F., & Ristuccia, L. (2026). Beyond military sales: The market premium on dual-use R&D in the defence sector. Economics Letters, 260, Article 112820.
• Bottai, C., de Rassenfosse, G., & Raiteri, E. (2025). A new approach to measuring invention commercialization: An application to the SBIR program. Research Policy, 54(9), Article 105302.
• Decker, J., & Li, E. (2023, July 14). How software companies can enter the U.S. defense market. Harvard Business Review.
• Global Affairs Canada. (2026). A guide to Canada's export control list: January 2026. Government of Canada.
• Innovation, Science and Economic Development Canada. (2026). State of Canada's defence industry 2026. Government of Canada.
• Markham, S. K., Ward, S. J., Aiman-Smith, L., & Kingon, A. I. (2010). The valley of death as context for role theory in product innovation. Journal of Product Innovation Management, 27(3), 402–417.
• Molas-Gallart, J. (1997). Which way to go? Defence technology and the diversity of 'dual-use' technology transfer. Research Policy, 26(3), 367–385.
• National Research Council Canada. (2026, January 9). Minister Joly announces over $240 million to boost defence innovation support for Canadian small and medium-sized businesses developing dual-use technologies. Government of Canada.
• Office of the Procurement Ombud. (2026). Turning complexity into capability: Canada's defence procurement system for major projects. Government of Canada.
• Prime Minister of Canada. (2026, February 17). Prime Minister Carney launches Canada's first Defence Industrial Strategy to strengthen security, create prosperity, and reinforce strategic autonomy. Government of Canada.
• Public Services and Procurement Canada. (2025). Guide to the Schedule to the Defence Production Act. Government of Canada.
• Public Services and Procurement Canada. (2026a). Register in the Controlled Goods Program. Government of Canada.
• Public Services and Procurement Canada. (2026b). How to meet Level 1 cyber security certification requirements. Government of Canada.
• Public Services and Procurement Canada. (2026c). Sponsorship process. Government of Canada.