AI Funding Programs in 2026 in Germany and Austria

In Germany and Austria, AI funding for 2026 is provided through several parallel programs that differ significantly in funding rates, maximum amounts, and application procedures. Anyone planning an AI project can quickly lose valuable time if they apply to the wrong program or budget for a funding amount that later turns out to be a cap on eligible costs. This article organizes the most important programs in both countries, explains the technical requirements for an application, and identifies the threshold at which a project becomes ineligible for funding. For companies planning to operate in both Germany and Austria, a third question arises: which funding pool corresponds to which location.

AI Funding in Germany in 2026

The research allowance under the Research Allowance Act is the most versatile tool. According to an overview by novaris-consulting from March 2026, it covers 25 to 35 percent of eligible personnel costs, with a maximum amount in the millions per year, and can be applied for retroactively. So, anyone who has already started an AI project can still apply for the grant retroactively.

According to the same overview, the ZIM and KMU-innovativ funding programs of the Federal Ministry of Research each cover a portion of the project costs, with no fixed percentage. Both programs are aimed at companies in industry and retail with a clearly defined research or development project. Funding rates and deadlines change over the course of a year; ultimately, only the information on the respective program website is binding.

AI Funding in Austria in 2026

The call AI Ökosysteme 2026 of the Austrian Research Promotion Agency (FFG) provides around 8.4 million euros, with 100,000 euros to 1 million euros per project. It funds collaborative industrial research projects in the focus areas Hybrid AI (program line AI for Tech) and Green AI (program line AI for Green), and each consortium needs at least two companies. The FFG requires the further development of core AI technologies; merely applying existing models to a use case is not enough. The submission deadline is October 6, 2026, and the application is written in English.

For a single development project without a set topic, the FFG Basisprogramm is the main route. It is open all year in 2026 and funds up to 3 million euros per project. Small and medium-sized enterprises and startups with smaller budgets use the Basisprogramm Kleinprojekt: a grant of up to 60 percent on total costs of no more than 150,000 euros, with a maximum of 88,500 euros in funding. The AI programs AI-Start and AI-Adoption of Austria Wirtschaftsservice (aws) currently have no open call. The AI-Start Green call closed on November 30, 2024, and the AI-Adoption call on February 28, 2025.

Independently of any call, companies can claim the Forschungsprämie (research premium). It amounts to 14 percent of the research expenses of a fiscal year, is credited by the tax office to the company's tax account and also benefits companies that make no profit. For in-house research, a free annual expert opinion from the FFG is required, which is requested via FinanzOnline after the end of the fiscal year. Companies with a site in Vienna can also use the funding programs of the Wirtschaftsagentur Wien.

As of September 2026. Budgets, funding rates and deadlines change constantly; only the current information from the respective funding body is binding. Check the conditions there before the project starts, because the Wirtschaftsagentur Wien and aws only fund projects that begin after the application has been submitted.

Companies operating in both countries must review the funding programs separately. A German subsidiary applies for the research tax credit through the German tax office, while an Austrian branch submits its application independently to aws or the FFG. Double funding of the same costs across both countries is not permitted; however, dividing the costs by location and task is common practice and is usually the simpler approach in practice.

What a grant application requires from a technical standpoint

Anyone applying for AI funding in 2026 will encounter the same fundamental questions in all of the programs mentioned, regardless of the funding rate. They require a technical risk whose solution is not yet clear at the start of the project, an approach that is new—at least to the applicant company—and a project that can be broken down into work packages with milestones.

In practice, the cost estimate often determines whether a proposal is approved or rejected. Personnel hours, material costs, and external services must be assignable to a specific work package, not just to the project as a whole. If you wait until you’re writing the proposal to establish this structure, you’ll waste time. From our own projects, we know that a rough cost estimate before the actual project begins can shorten the later grant application process by several weeks.

The most common reason for rejection is a technical risk that is described too vaguely. Anyone who writes that a system is intended to increase efficiency will receive follow-up questions from the funding agency. On the other hand, those who describe why an existing method fails when applied to their own data—and explain that the solution is initially uncertain—have a better chance of receiving approval.

From Proposal to Pilot Project

Once funding has been approved, the actual work on the use case begins. Our „90-Day AI Pilot Project Methodology“ shows how funding can be used to develop a prototype with real users, rather than ending up as just a presentation. Anyone who wants to assess technical feasibility before submitting an application will find guidance on this in a Initial Consultation on AI Consulting von torck offers a brief, no-obligation introduction. The article shows which initial projects are suitable for getting started Generative AI in Industry and Retail: Three Entry Projects with Measurable Benefit.

If the project involves running your own language model, the hardware belongs in the budget. Which model size at which quantisation fits a given GPU is shown by the LLM Finder with the VRAM requirement for each configuration.

Frequently Asked Questions

Which program is right for which project?

For a single company with a clearly defined research component, the research allowance (Forschungszulage) is the easiest way to get started because it does not involve a competitive selection process. ZIM and KMU-innovativ are suitable for projects involving collaborative partners. In Austria, the FFG Basisprogramm or the Basisprogramm Kleinprojekt suits a single development project, while the FFG call AI Ökosysteme suits consortia that advance core AI technologies. Like the research allowance, the Forschungsprämie is claimed through the tax office. If you’re unsure, it’s best to clarify the classification before writing the application, not after.

How long does it take to process an application?

Depending on the program, it can take anywhere from several weeks to a few months between the initial proposal and approval. The research grant is processed through the tax office and requires a technical certification; this is usually faster than a traditional review process involving a jury. For FFG calls for proposals with a fixed deadline, the next submission deadline sets the pace, not your own project start date. If you submit early enough, you’ll have time to address any questions from the funding agency without having to postpone the project’s start date.

Can you combine multiple programs?

In principle, yes, as long as the same costs are not funded twice. The research allowance can be combined with other German programs, with the amounts credited toward the total. In Austria, tax-exempt grants, such as those from the FFG, reduce the assessment basis of the Forschungsprämie. For de minimis funding, such as that of the Wirtschaftsagentur Wien, all de minimis aid counts together toward the ceiling. The specific conditions for combining grants are ultimately clarified by the respective funding agency before the application is submitted.

What AI funding programs are available in Austria?

As of September 2026, the main options open for AI projects are the FFG Basisprogramm and the Basisprogramm Kleinprojekt, plus the FFG call AI Ökosysteme 2026 with a submission deadline of October 6, 2026, and the Forschungsprämie of 14 percent of research expenses. SMEs in Vienna can also use the Förderung Innovation and the Förderung Digitalisierung of the Wirtschaftsagentur Wien. AI-Start and AI-Adoption of aws currently have no open call, and KMU.DIGITAL no longer accepts applications.

Why torck

torck develops software with its own teams in Maxhütte-Haidhof, Vienna, and Rabat, and supports AI projects from the proposal stage through to ongoing operations. Whether you submit an application in Germany or a second one in Austria, you’ll be working with the same development teams and the same contractual partner—the German company torck GmbH—rather than a chain of subcontractors. We handle the technical implementation of funded projects and subsequently operate them as well. If you want to know which program is right for you and what the implementation will actually cost, the best way to find out is through a Initial Consultation on AI Consulting.

Schedule a meeting