The one rule that governs all stacking
Every legal stack in the Canadian funding landscape follows a single principle: you cannot claim the same dollar twice. No program prohibits you from receiving funding from multiple sources — they prohibit you from recovering the same specific expenditure more than once.
In practice this means programs stack by covering different cost categories, different cost pools, or the non-funded portion of a shared project. IRAP covers a defined portion of eligible labour. SR&ED then applies to the eligible labour IRAP didn't cover, plus materials and overhead. FedDev covers commercialization and capital costs that neither IRAP nor SR&ED touches. Each program occupies a different slice of your total spend.
The core mechanic
Programs stack because they cover different things. IRAP is a labour grant. SR&ED is a tax credit on R&D expenditures. FedDev is a scale-up loan. Ontario OITC is a provincial top-up on SR&ED. Scale AI covers project costs for consortium-based deployments. None of these are the same instrument applied to the same cost — which is exactly what makes combining them legal.
Which programs can be combined
IRAP + SR&ED + Ontario OITC
The most widely used combination for early-stage Canadian AI companies. IRAP covers a portion of labour as a monthly grant; SR&ED credits the remaining eligible labour at 35% at year-end; OITC adds 8% on top of SR&ED for Ontario teams. Each covers a distinct cost pool with no overlap.
SR&ED + FedDev Ontario BSP
SR&ED covers R&D expenditures; FedDev covers commercialization and scale-up costs. Different program mandates mean they naturally target different budget lines. Commonly used by Southern Ontario companies with both active R&D and commercial expansion underway.
Scale AI + IRAP + SR&ED
For consortium-based applied AI projects. Scale AI covers 40% of total project costs; IRAP covers internal R&D labour not funded by Scale AI; SR&ED applies to the eligible experimental development component. Requires careful cost allocation between the three programs but produces the highest total recovery rates.
IRAP + Scale AI
Possible but requires clear separation. Scale AI funds the consortium project costs; IRAP funds internal labour on R&D work that sits outside the Scale AI contribution scope. The programs can't fund the same labour pool — an ITA will help structure the split if you're pursuing both.
FedDev + IRAP
Sequential more than simultaneous. IRAP typically funds R&D during development; FedDev funds the subsequent scale-up phase. Different cost categories apply to each, but the timing and project scope usually mean they serve different chapters of the same company's story rather than running in parallel.
ACAF + IRAP + SR&ED
When open, ACAF covered GPU and cloud compute costs — a category neither IRAP nor SR&ED covers directly. Adding ACAF to the standard IRAP + SR&ED stack significantly boosted recovery for compute-heavy companies. The intake is closed as of June 2026. Monitor ised-isde.canada.ca for a second round.
Why the cost category separation matters
The reason stacking works is that the five programs were designed with different mandates — they were never intended to be mutually exclusive. SR&ED was created to incentivize R&D through the tax system. IRAP was created to provide direct support for SME innovation. Scale AI was created to accelerate applied AI in specific sectors. FedDev was created to drive regional economic development. ACAF was created to address a specific compute access barrier.
When your company spends money on R&D and commercialization, that spend naturally falls across multiple categories: labour, compute, equipment, marketing, capital. Because each program targets different categories, a company with diverse spend can legitimately access multiple programs — as long as the accounting is clean and no cost is reported under more than one program's eligible expenditure base.
Where stacking goes wrong
The most common mistake is claiming the same labour costs under both IRAP and SR&ED. If IRAP funded $187,500 of $250K labour, that $187,500 is excluded from your SR&ED eligible expenditure calculation. The SR&ED claim applies only to the remaining eligible R&D spend not covered by government assistance. Claiming both programs on the same costs — even accidentally — triggers CRA review and repayment obligations.
Which stacks apply to which company types
Pre-revenue AI startup — Ontario CCPC
Building novel AI technology, under 500 employees, active R&D spend, Toronto or Waterloo base
IRAP
SR&ED
OITC
Revenue-generating Ontario scale-up
Scaling an AI product, 2+ years revenue, audited financials, active commercialization
FedDev BSP
SR&ED
OITC
Manufacturer adopting AI — Southern Ontario
Established manufacturer implementing AI quality control, logistics, or predictive maintenance with an AI tech partner
Scale AI
IRAP
SR&ED
Company with tariff exposure — Southern Ontario
25%+ of sales to US or China, documented cost impact from trade measures, active R&D ongoing
FedDev RTRI
SR&ED
IRAP
Compute-heavy AI company (when ACAF reopens)
Significant GPU or cloud spend alongside active R&D labour — model training, inference infrastructure
IRAP
SR&ED
ACAF*
* ACAF intake closed as of June 2026. Monitor ised-isde.canada.ca for reopening.
The sequencing question
Which program to pursue first matters almost as much as which programs to combine. IRAP requires an ITA relationship to be in place before R&D spending begins — you can't retroactively claim it. SR&ED can be claimed on work already done within 18 months of fiscal year-end. Scale AI requires a consortium and an EOI before you spend. FedDev is best pursued once commercial momentum is established.
For most early-stage AI companies, the right sequence is: engage an ITA for IRAP as early as possible, begin SR&ED documentation from day one, and assess Scale AI and FedDev once the company profile fits those programs' requirements. The programs aren't mutually exclusive in timing — a company can be active in IRAP, filing SR&ED annually, and exploring a Scale AI consortium simultaneously.