The Forschungszulage doesn't fund every piece of technical work, only projects that meet the definition of qualifying research and development under § 2 FZulG. Whether that applies to your project is decided by the Bescheinigungsstelle Forschungszulage (BSFZ) against the Frascati criteria from the OECD Frascati Manual. Three characteristics have to be present at the same time. If one is missing, there is no certificate, and so no allowance.
The three R&D categories under § 2 FZulG
The Act recognises three qualifying categories. All three are eligible; the category doesn't affect the amount of funding, but it does shape the argument in the BSFZ application.
- Basic research — experimental or theoretical work aimed primarily at new knowledge about the underlying nature of phenomena and observable facts, with no specific application in view.
- Industrial research — planned research or critical investigation aimed at acquiring new knowledge and skills that can be used to develop new products, processes or services.
- Experimental development — acquiring, combining, shaping and using existing scientific, technical and other knowledge and skills with the aim of developing new or substantially improved products, processes or services. This includes prototypes, pilot projects, demonstrations and first trials.
Most corporate R&D projects fall into the experimental development category. That is no disadvantage: the funding rate and assessment base are identical for all three.
The Frascati criteria: novelty, uncertainty, systematic method
The BSFZ's assessment guide (as at October 2025) operationalises three overarching characteristics derived from the OECD Frascati Manual 2015. All three have to be evidenced simultaneously.
- Novelty — the project seeks new knowledge or new solutions. New means not already state of the art at the time the project started. The relevant yardstick is world knowledge in the field, not your company's state of knowledge. Introducing a method that is new internally but known externally doesn't meet the novelty test.
- Scientific and technical uncertainty — the outcome of the project was not foreseeable when it started. The goal is known, but the route to it, or whether it is technically achievable, is open. What is required is a genuine knowledge problem, not an implementation risk.
- Systematic method — the project follows a planned methodology with comprehensible work steps, milestones and a research logic. Unplanned trial and error without documentation generally doesn't meet this test.
A fourth characteristic appears in the Frascati Manual but isn't treated by the BSFZ as a separate test: the transferability of results. Even so, it does no harm to describe in the application whether and how project results can be generalised; it strengthens the novelty argument.
What doesn't count as R&D
The BSFZ rejects applications where the project meets none of the three characteristics, or where it is one of the following activities that the Frascati Manual expressly excludes.
- Routine product development — adapting existing products to customer wishes, new colours, variants or versions with no new technical knowledge.
- Quality assurance and testing — ongoing quality checks and standard tests in production, even where they are technically demanding.
- Market research and customer studies — even where carried out systematically.
- Introducing known technologies — migrating to a new platform, rolling out standard software or digitalising existing processes with no new scientific or technical content.
- Project management and administration — coordination tasks, even where they are assigned to an R&D project.
What matters isn't the topic but the substance. A project to digitalise production can be eligible if novel algorithms are developed in the process whose functioning is uncertain at the outset. The same topic without technical uncertainty is not eligible.
Drawing the line in practice
The most common rejection at the BSFZ arises not from technical triviality but from a wrong description. Anyone describing the market benefit or the customer advantage in the application, instead of evidencing the technical challenge and the advance in knowledge over the state of the art, fails at least two of the three tests.
- Not sufficient: "We are developing an AI-supported platform for automated quality inspection that saves our customers 30 per cent of their time." — No indication of technical uncertainty or novelty over the state of the art.
- Sufficient: "We are investigating whether transformer models can be used to detect faults in high-frequency sensor data when the training basis is under 500 labelled examples. Technical feasibility and the accuracy threshold are open at the start of the project." — A clear statement of the scientific question, the knowledge problem and the uncertainty.
Projects containing several components — R&D elements alongside routine elements — have to be cleanly separated. The BSFZ certifies only the R&D elements; mixed projects with no delimitation risk being rejected in full, even where part would have been eligible.
How the BSFZ assesses
The BSFZ assesses only whether the project qualifies on the merits. Not the level of costs, not staff records. The application is made online through the portal bescheinigung-forschungszulage.de. Processing typically takes two to four months.
The assessment is based on the project description submitted. That description should cover the following points:
- The state of the art immediately before the project started: what was already known?
- The specific research question or technical challenge: what is still unknown?
- Why is the solution technically uncertain at the start of the project?
- The methodological approach and a work plan with milestones.
- The qualifications of the staff involved (optional, but helpful).
The BSFZ certificate is a basic assessment notice under § 171 (10) AO and binds the tax office on the question of whether the project qualifies. At the second step the tax office independently assesses the assessment base: the level of eligible staff costs, own work and contract research. A BSFZ certificate is a necessary but not sufficient condition for the Forschungszulage.
A full overview of the funding rate, the assessment base and the two-stage procedure is in the guide Forschungszulage 2026.
FAQ
What are the Frascati criteria for the Forschungszulage? The BSFZ assesses every project against three characteristics from the OECD Frascati Manual: novelty, scientific and technical uncertainty, and systematic method. If one of those three is missing, the certificate is refused.
Does product development count as eligible research? Not automatically. Product development is only eligible where it overcomes genuine technical uncertainty and shows novelty within the meaning of the Frascati Manual. Routine product development that transfers known methods to new applications without generating new knowledge is not eligible.
What is the difference between industrial research and experimental development? Industrial research aims at new knowledge and new capabilities, for instance investigations that don't yet have a specific product in view. Experimental development uses existing knowledge and builds new prototypes, pilot plants or first systems on it. Both categories are eligible under § 2 FZulG.
How do I describe my project correctly for the BSFZ? Describe the state of the art before the project started, the technical challenge with no known solution, and the methodological route to a solution. Market benefit or competitive advantage are irrelevant to the BSFZ's assessment: it examines the technical and scientific substance, not the business value.
Is software development eligible R&D under the FZulG? Yes, where the development aims at a technical and scientific advance that is uncertain at the outset. Pure feature development, bug fixing or adapting existing libraries without new knowledge don't count as R&D. Novel algorithms, compiler optimisations or machine learning with an open outcome, on the other hand, can meet all three Frascati criteria.
Forschungszulage 2026: the complete guide
A €12m assessment base, a 35% SME rate, a two-stage procedure: an overview of the Forschungszulage in 2026, and where to find the detail.
The BSFZ certificate: process, criteria and processing time
No Forschungszulage without a BSFZ certificate. What the certification body assesses, how the application works, and why many descriptions start from the wrong point.
Forschungszulage for software development and AI: what counts and what doesn't
Software development and AI are eligible, provided novelty, uncertainty and a systematic method can be evidenced. What that means in concrete terms and where applications fail.
Forschungszulage rejected: reasons and what you can do
The BSFZ or the tax office has rejected your application. These are the most common reasons and your concrete options: objection, new application, revision.
All the detail, eligibility and pitfalls:
- [1]§ 2 FZulG — qualifying research and development projects— Federal Ministry of Justice · gesetze-im-internet.de, 2026 Source
- [2]Frascati Manual 2015 — Guidelines for Collecting and Reporting Data on Research and Experimental Development— OECD, 2015 Source
- [3]Assessment guide of the Bescheinigungsstelle Forschungszulage (as at October 2025)— Bescheinigungsstelle Forschungszulage (BSFZ), 2025 Source
- [4]FAQ — assessment criteria and application procedure for the Forschungszulage— BSFZ, 2026 Source
- [5]Research Allowance Act (FZulG) — consolidated version— Federal Ministry of Justice · gesetze-im-internet.de, 2026 Source