Business Finland

How much Business Finland R&D funding can an AI project receive?

Funding ranges for AI R&D projects by company size and cost category, including cost-share levels, personnel cost logic, and typical project sizes.

There is no single number. Business Finland R&D funding is a percentage of your eligible project costs, so the amount you receive is a function of two things: how large and credible your project budget is, and what funding rate applies to your company and project type. Understanding that formula is more useful than any headline figure, because it tells you how to size a project that is worth funding.

The specific rates and caps below move over time and depend on your situation. Treat them as the shape of the model, and confirm current figures for your case before you budget against them.

The formula

Your funding is roughly:

eligible costs × funding rate = funding amount

If a project has EUR 300,000 of eligible costs and the applicable rate is 50%, the funding is around EUR 150,000. The rest is your own share, which you finance yourself. This is why financial standing matters so much — you are always co-investing.

Funding rates

The rate depends on company size and how the funding is structured:

Grant share for R&D commonly lands around 50% of eligible costs for company development projects, with higher rates (often cited near 65%) available for more research-intensive work or under specific instruments and collaboration setups.
SMEs generally access more favourable treatment than large companies.
R&D loans can cover a larger share of costs — often up to around 70% — because they are repayable. Many larger projects combine a grant portion with a loan portion.

The practical takeaway: a research-heavy AI project with real uncertainty can attract a higher grant rate than a project that looks close to a commercial build.

What counts as eligible cost

The funding amount is built from the cost base, so the cost base is where you should focus. Eligible categories typically include:

Personnel costs — salaries for the people doing the R&D work, based on actual hours booked to the project. This is usually the largest line for an AI project.
Indirect costs / overhead — an allowance calculated as a percentage on top of personnel, covering general expenses and personnel side costs rather than being itemised.
Purchased services and subcontracting — external research, specialist development, or expertise you buy in, usually capped as a proportion of the project.
Materials, equipment, and travel — direct project costs, including relevant depreciation on equipment.

Costs that are not clearly attributable to the R&D work, or that look like normal business-as-usual, get stripped out during assessment and shrink your funding.

Typical AI project sizes

In practice, credible AI R&D projects tend to sit in the EUR 100,000–500,000+ range of total costs, translating to funding somewhere from the low six figures upward. Smaller than that and the project often looks too minor to be a serious R&D bet; much larger and the assessment scrutiny (and your own co-financing burden) rises sharply.

A useful anchor: a project built mostly around a small senior team working for several months to a year, plus some specialist subcontracting, naturally lands in this band once personnel and overhead are counted properly.

Why “as much as possible” is the wrong goal

Applicants often try to maximise the funding number. Evaluators read an inflated budget as a signal that the project logic is loose. A budget that is tightly tied to work packages, staffed realistically, and honest about subcontracting reads as credible — and credible projects get funded. An oversized budget with a hand-wavy plan invites cuts or rejection.

The right question is not “what is the maximum we can ask for?” It is “what does this specific development work genuinely cost, and how does each euro map to a work package?” Answer that well and the funding amount takes care of itself.

Before you model the budget

Lock the project scope and work packages first, then cost them. Building the budget before the plan is the most common way companies end up with a number they cannot defend in assessment.