SR&ED Tax Credits for Startups: What Qualifies and How to Claim

For a Canadian startup doing real technical work, the Scientific Research and Experimental Development (SR&ED) program is often the single largest source of non-dilutive money available. Eligible small Canadian-controlled companies can get a substantial share of their qualifying R&D spending back as a cash refund, even if they have no taxable income. Yet plenty of founders either don't claim or claim badly.

This guide explains how SR&ED works in plain language, what qualifies, how to document it and how to avoid the problems that lead to reduced claims. Rates, limits and administrative rules change, so confirm current details on the Canada Revenue Agency (CRA) website and with an experienced SR&ED adviser before you file.

What SR&ED is

SR&ED is a federal tax incentive administered by the CRA. It rewards Canadian businesses for doing systematic research or development work that tries to advance technology. It works through investment tax credits (ITCs) calculated on qualifying expenditures.

Two rates matter most:

• Enhanced rate for eligible CCPCs: Canadian-controlled private corporations can earn a 35% refundable credit on qualifying expenditures up to an annual expenditure limit. "Refundable" means you get cash back even if you owe no tax.
• Basic rate: 15%, generally non-refundable, for other corporations and for spending above the limit.

The expenditure limit and how it phases out with company size have been subject to recent federal proposals and changes, so check the current figures. Most provinces also offer their own R&D credits on top of the federal program, some of them refundable, so the combined benefit can be higher.

What work qualifies

This is where most confusion lies. SR&ED isn't about whether your product is innovative in a business sense. It's about whether the work involved trying to overcome technological uncertainty through a systematic process.

The CRA uses questions along these lines to assess eligibility:

1. Technological uncertainty: was there a technical problem that couldn't be solved using standard practice or readily available knowledge?
2. Hypotheses: did you formulate ideas specifically aimed at reducing that uncertainty?
3. Systematic investigation: did you test those ideas through experimentation or analysis, in a structured way?
4. Technological advancement: was the goal to advance your understanding of the underlying technology, whether or not you succeeded?
5. Records: did you keep records of the hypotheses, tests and results as you went?

Failure counts. If you ran experiments that didn't work, that work can still qualify, provided it was systematic and aimed at a technological advancement.

Examples that often qualify

• Developing a new algorithm because existing approaches couldn't meet performance or accuracy requirements, with documented experiments.
• Building a system architecture to handle a scale or latency constraint that standard tools couldn't meet.
• Creating new materials, processes or devices where the outcome was uncertain.
• Developing machine learning models where off-the-shelf methods failed and new approaches had to be tested systematically.

Work that usually doesn't qualify

• Routine software development using known techniques, such as standard web or mobile app features.
• User interface and visual design changes.
• Market research, sales promotion and business strategy.
• Routine testing and quality control.
• Social science and humanities research.
• Data collection that isn't part of a systematic investigation.

Most software startups have a mix. A product might involve 20% to 40% genuinely uncertain technical work and the rest routine development. The claim should only include the qualifying part.

Which costs qualify

Qualifying expenditures typically include:

There are special rules on specified employees (owners and related people), on payments to non-arm's-length parties and on government assistance. If you received an IRAP grant for the same work, for example, it generally reduces your eligible SR&ED expenditures. Work done outside Canada generally doesn't qualify.

How to document your work

The single biggest reason claims get reduced is poor documentation. The CRA wants evidence created at the time the work was done, not a story written a year later.

Good habits:

• Use the tools you already have. Issue trackers, pull requests, design documents, experiment logs and notebooks are all useful evidence if they describe what you were trying and why.
• Tag SR&ED-relevant work. A simple label in your ticketing system makes it much easier to identify qualifying projects later.
• Track time. Even rough weekly allocations by project are far better than estimates made after the fact.
• Record the uncertainty. When starting a hard technical task, write a sentence on why standard approaches won't work and what you'll try.
• Record the results. Note what happened, including failures, and what you concluded.

A note like that, written at the time, does much of the work of supporting a claim.

How to file

SR&ED is claimed with your corporate income tax return. The main forms are the T661 (describing the projects and expenditures) and Schedule 31 (calculating the investment tax credit). The deadline is generally 18 months after the end of the tax year in which the expenditures were incurred. Miss it and you lose the claim for that year.

The technical narrative on the T661 is critical. For each project, you describe the technological uncertainty, the work done to address it and the advancement achieved, within word limits. Write it in clear, specific language. Avoid marketing phrases. Reviewers are often scientists or engineers and respond best to concrete technical detail.

Reviews and how to handle them

The CRA may review a claim, particularly a first claim. A review can involve a financial reviewer checking the costs and a research and technology advisor checking the technical eligibility. Reviews can feel stressful but are a normal part of the program.

Be prepared to walk the reviewer through your projects, show contemporaneous documentation and explain your time allocations. Companies with good records usually come through reviews with their claims intact or only slightly adjusted.

Using a consultant

Many startups hire SR&ED consultants, who often charge a percentage of the credit. A good consultant can help identify eligible work, write strong narratives and handle reviews. A poor one can over-claim aggressively and leave you exposed.

Questions to ask:

• How do you decide which work is eligible, and will you push back if something isn't?
• What's your fee, and is there a cap?
• Will you support us through a CRA review at no extra cost?
• Can we speak to two or three clients in our sector?

Some startups use their accountant for the financial side and write the technical narrative themselves. That works well when the technical team is strong at clear writing.

SR&ED and fundraising

Investors in Canadian startups know SR&ED well, and many expect you to claim it. Build expected refunds into your financial model, but be conservative about timing. Refunds can take months to arrive, especially for first-time claims or if there's a review.

Be careful about anything that might affect your CCPC status, such as taking a controlling investment from non-residents. Losing CCPC status can reduce the credit from the enhanced refundable rate to the basic non-refundable rate. Our guide on raising money in Canada covers this.

Financing your SR&ED claim

Because refunds arrive after year-end, some lenders offer SR&ED financing: a loan against your expected refund. It can bridge a cash gap, but fees and interest vary. Compare offers carefully and only borrow against claims you're confident in.