Canada rewrote its largest innovation subsidy. Most founders are still budgeting for the old rules

Every discussion of Canadian innovation policy eventually arrives at the Scientific Research and Experimental Development program, usually called SR&ED and pronounced as a word. It is by a wide margin the largest federal support for business research and development, delivering billions of dollars a year in tax credits to tens of thousands of claimants, and for a large share of Canadian technology companies it is the single biggest cheque the government writes them.

It is also, by long-standing consensus, administratively difficult, unevenly applied and poorly understood by the companies it is meant to serve. The 2024 Fall Economic Statement proposed the most substantial changes to it in years, and the practical effect is concentrated in a stage of company that rarely has a tax specialist on staff.

Start with how the program actually works, because the structure explains why the changes matter. SR&ED provides a tax credit on eligible expenditures, principally the salaries of people doing qualifying work. A Canadian-controlled private corporation can access an enhanced credit rate of 35 percent, and critically that enhanced credit is refundable, meaning the company receives cash even if it owes no tax. For a pre-revenue company, refundability is not a detail. It is the entire value of the program. A non-refundable credit is worth nothing to a firm with no profit to offset.

The enhanced rate applies only up to an annual expenditure limit. Above that limit, the credit rate falls to the general rate, which is lower and, for most claimants, non-refundable. The expenditure limit is therefore the ceiling on how much genuinely useful support a growing company can get, and it had been set at $3 million for a long stretch during which salaries did not stay still.

The reform raises that limit to $4.5 million. At the 35 percent enhanced rate, that moves the maximum refundable credit from roughly $1.05 million to roughly $1.575 million a year. For a company running a research team of thirty or forty people, that difference is a meaningful fraction of a funding round, and it arrives without dilution.

The second change is less visible and affects more companies than the first. The expenditure limit has always been ground down as a company grows, phasing out based on taxable capital employed in Canada. The old range was $10 million to $50 million, meaning a company with $50 million of taxable capital lost access to the enhanced rate entirely. The reform moves that range to $15 million to $75 million. In plain terms, companies stay in the enhanced regime substantially longer, and the point at which the program stops being useful shifts later into their growth.

The third change extends eligibility for the enhanced refundable credit to Canadian public corporations. This is a departure from the program's historical logic, which restricted the enhanced treatment to Canadian-controlled private corporations on the theory that public companies had access to capital markets and needed less help. The reasoning was never especially strong for small listed companies, many of which are capital-constrained in exactly the way the program is designed to address, and a listing on a Canadian exchange had the perverse effect of reducing a company's support.

There is also a proposal to restore eligibility for capital expenditures, which had been removed in 2014. That matters unevenly. A software company spends its research budget on salaries and barely notices. A company building hardware, instruments, laboratory equipment or pilot production lines spends heavily on capital, and its effective support rate had quietly fallen for a decade. Restoring capital eligibility disproportionately benefits the sectors Canada most often says it wants: advanced manufacturing, clean technology, medical devices, semiconductors.

What the reform does not address is the part founders complain about most, which is the experience of claiming. SR&ED is self-assessed and reviewed after the fact, and eligibility turns on whether work constituted systematic investigation to resolve technological uncertainty. That is a real distinction with a coherent basis: the program funds attempts to solve problems whose solutions were not knowable in advance, not ordinary engineering. It is also a judgement call, and judgement calls made in review, months or years after the work, are where the uncertainty lives.

The consequence is an entire advisory industry that exists to prepare claims, frequently on contingency at a substantial share of the recovery. Money that Parliament appropriated to fund research ends up funding the documentation of research. That is a deadweight cost, it falls hardest on the smallest claimants who can least afford specialists, and no adjustment to rates or thresholds touches it.

The timing problem is similar in character. A company incurs salary costs through the year, files after year end, and receives the credit some months later. The gap between spending and receiving can approach eighteen months, which for a company managing a finite runway is a working capital problem regardless of the eventual amount. A specialised lending market has grown up to bridge exactly this gap, which is a reasonable market response and also a sign that the instrument is not delivering support when it is most needed.

It is worth setting this against what other countries have done, because Canadian firms compete for capital internationally. The United States moved in the opposite direction for several years by requiring research costs to be amortised rather than deducted immediately, which materially worsened cash flow for American startups before it was addressed. The United Kingdom tightened its equivalent scheme after significant abuse. Against that backdrop, a Canadian reform that raises limits and broadens eligibility is a genuine competitive improvement, and it is the kind of thing that rarely gets mentioned in the pitch for moving a company south.

The practical advice for a Canadian technology company is narrow and unromantic. Establish whether your taxable capital sits inside the new phase-out range, because that determines whether any of this reaches you. Model the higher expenditure limit against next year's research payroll rather than last year's. If you are a listed company that stopped claiming the enhanced credit because you were not eligible, check again. And document contemporaneously rather than reconstructing at year end, because the reform changed the arithmetic of the program and left the evidentiary standard exactly where it was.