Why most AI companies only find one program
The five programs share no common application, no coordinating body, and no shared eligibility language. SR&ED is administered by CRA. IRAP runs through NRC. Scale AI is a separate industry consortium. FedDev is a regional development agency. Each has its own intake process, its own reviewers, and its own definition of what qualifies.
Most founders discover one — usually SR&ED through an accountant or IRAP through a peer referral — and assume that's the landscape. The others go unclaimed. This page is the map.
The five programs
Canada's largest R&D program, distributing $4.5B+ annually. CCPCs receive a 35% refundable tax credit on eligible R&D expenditures — meaning CRA sends a cheque regardless of profitability. The Bill C-15 expansion in March 2026 doubled the annual expenditure limit to $6M, raising the maximum refundable credit to $2.1M per year. Pre-revenue companies qualify. Ontario adds a further 8% refundable credit (OITC) on top.
Canada's flagship non-repayable R&D grant, covering up to 75% of eligible Canadian labour costs. Unlike SR&ED, IRAP pays monthly via reimbursement — better for runway than a year-end tax credit. The AI Assist sub-stream is specifically for SMEs adopting or building AI, with no AI background required. Applications flow through an Industrial Technology Advisor (ITA) assigned to your region — there's no open portal. First-time applicants typically receive $75K–$500K.
Canada's AI Supercluster, funded through the federal Innovation Supercluster Initiative. Co-invests in applied AI projects at 40% of eligible costs (50% for Quebec-based projects). As of March 2025, Scale AI has supported 162 projects across 630+ organizations with a total portfolio of CAD $673M. The program requires a consortium of at least two Canadian companies — a solo application is automatically rejected without review. Supply chain AI, healthcare, manufacturing, and logistics consistently score well.
The federal regional development agency for Southern Ontario operates two active streams. The Business Scale-up and Productivity (BSP) stream provides interest-free repayable loans for scaling companies. The Regional Tariff Response Initiative (RTRI) is non-repayable for businesses with documented tariff exposure. Both require companies to be located in Southern Ontario — roughly the region south of the French River. Rolling intake with a 2–4 month decision timeline.
A $300M fund covering GPU and cloud AI compute costs, announced as part of Canada's National AI Strategy. The only Call for Proposals closed July 31, 2025. No second round has been announced as of June 2026. Canadian cloud providers (OVHcloud Canada, Borealis AI, NVIDIA DGX Cloud Canada) received 66% coverage vs. 50% for US hyperscalers — a meaningful difference at scale. Worth monitoring if you have significant compute spend.
Quick comparison
| Program | Type | Max amount | Revenue req. | Consortium | Ontario only | Intake |
|---|---|---|---|---|---|---|
| SR&ED | Tax credit | $2.1M/yr | None | No | No | Rolling |
| NRC IRAP | Non-repayable | $10M | None | No | No | Rolling |
| Scale AI | Non-repayable | ~$1.2M avg | None | Required | No | Ongoing |
| FedDev Ontario | Loan / grant | $10M | Must fund 50% | No | Yes | Rolling |
| ACAF | Non-repayable | Part of $300M | None | No | No | Closed |
Which program to start with
For most early-stage Canadian AI companies, the answer is SR&ED first and IRAP in parallel — they have the lowest barriers, the widest eligibility, and the most direct path to cash. SR&ED requires no relationship, no proposal, and no intake window — you file with your T2. IRAP requires an ITA relationship but pays monthly, which is better for runway.
Scale AI is worth pursuing if you have or can form a consortium and your project maps to their priority sectors. FedDev is underused by Toronto-area founders who don't realise they qualify simply by being in Southern Ontario. ACAF should be on your radar if you have significant GPU or cloud compute spend and a new intake round is announced.
The order and combination that makes the most sense depends on your company's size, location, funding stage, and what technical work is currently underway. The programs that appear straightforward on paper often have eligibility nuances — and the ones that appear complex often have clearer paths than they look.