State grants for technology companies in Russia are more accessible than most founders assume. The Foundation for Innovation Assistance (Фонд содействия инновациям) runs competitive programs for early and growth-stage tech businesses. clickfraud.ru received a grant of approximately 8M RUB. Here's what the process actually looked like.
Why we applied
We weren't in financial trouble. We applied because the grant offered non-dilutive capital to accelerate product development — specifically, building out the machine learning layer of our click fraud detection system. Equity-free funding for something we were already planning to build was an easy decision.
The application structure
The Foundation's applications require several components that are worth thinking about carefully:
- Market sizing — you need to demonstrate addressable market with sourced data, not optimistic back-of-napkin math. We used data from Yandex and Google's public reports on Russian digital advertising spend, combined with industry estimates on fraud rates.
- Technical differentiation — what makes your approach novel? For us, this was the combination of behavioral fingerprinting with advertiser-side integration, which is genuinely different from IP-blocklist approaches.
- Team credentials — engineering background matters here. Our years of operating in digital advertising gave us real credibility.
- Financial projections — these need to be internally consistent and defensible. We modeled three scenarios (base, optimistic, conservative) with explicit assumptions.
The defense
The expert panel asked hard questions about scalability and the legal basis for collecting behavioral data. We'd anticipated both. The scalability question was answered with architecture diagrams. The legal question was answered by a prepared memo on Russian data processing law under 152-ФЗ.
One thing that helped significantly: we could demonstrate actual revenue and existing clients. This is not a requirement for the grant, but it's a powerful signal that you're not pitching a concept — you're funding an operating business.
What we used the funds for
The grant funded approximately 18 months of engineering work on the ML layer. Specifically: training infrastructure, data labeling, model evaluation pipelines, and the integration work required to deploy the new detection system alongside the rule-based system we already had.
Advice for other founders
- Apply early — the programs have competitive windows and you will likely need to apply more than once to understand what they're looking for.
- Be concrete about what the money will fund — "product development" isn't enough; specific engineering deliverables are.
- If you have revenue, show it. It removes the biggest risk question from the room before you walk in.
- The Foundation's website has detailed materials on what past successful applications looked like. Use them.
Grants are not free money — they come with reporting requirements and milestones. But they're far cheaper than equity, and they force you to articulate your roadmap with unusual precision.
— notes after receiving the grant, 2019
