Canada has great science and struggles to turn it into economic value. Our higher-education R&D spending ranks above the G7 average. Our business R&D is among the lowest in the G7 and has been falling for two decades. That is the commercialization gap, and with economic sovereignty under pressure it matters more than it did. Last week the government named the first 64 Eddie Goldenberg Research Chairs — $504M over eight years, at 21 universities. One of five selection criteria covers knowledge translation, which includes but is not limited to commercialization. So I asked a simple question: did it work? What commercial track record did these 64 people actually bring with them? 41% had already commercialized their science before they were selected — generating intellectual property, founding a company, or holding research-linked equity. 13 had founded 29 companies between them, and 9 of those were serial founders. One founded a US company 24 years ago now worth more than US$30 billion. That is about 15 points ahead of the most comparable earlier programme, the Canada Excellence Research Chairs, which had no separately weighted translation criterion. The samples are small — 64 and 34 — so neither difference reaches significance. The direction is consistent on both measures and suggestive, not established. Two things surprised me. First, I could find no equivalent analysis anywhere. Not in Europe, the UK, the US, Australia, Japan, Korea or the Nordics. Funders do publish commercialization numbers — the ERC found 44% of surveyed patent-holding grantees had commercially used their patents — but those are all ex post: what researchers did with the money. Nobody appears to check what a newly announced cohort had already done before it was chosen, which is the question that actually tests a selection criterion. Second, incidence is measurable and value mostly is not. 21 of the 29 companies carry no public figure of any kind. Private rounds go unreported, licensing terms are confidential. We can say whether commercialization happened. We mostly cannot say what it was worth. I used a deliberately inclusive definition. A patent counts on its existence, whether or not it was ever licensed — that accounts for seven of the 26. On the strictest reading, requiring evidence something was actually exploited, the figure is 25% rather than 41%. Both numbers are in the piece. My ask is modest. The criterion already exists; weight commercialization more heavily within it. And weight repeat founders and significant company-builders most, because those are the people most likely to shift the culture. The methods note and the full spreadsheet are published. Every coding decision carries its evidence and its source, and the counts on the summary sheet are formulas over the rows — so if you disagree with a call, change it and watch what moves. Corrections welcome; I will update the file. Let’s aim our funding programmes more squarely at the challenge we actually have: turning science into economic value. Link in first reply.