AI Funding DACH Mid-Market 2026: 7 Active Programmes + Workflow

Digital Jetzt and go-digital are dead. QCG covers up to 100% of AI training costs — almost nobody knows. 7 active funding programmes 2026 with application workflow for DACH mid-market.

Key numbers at a glance

Digital Jetzt expired on 31.12.2023, go-digital expired on 31.12.2024 (BMWK confirmation early 2025). Yet thousands of mid-market firms still search for these programmes in 2026 — wasting time on non-existent applications.
QCG (Qualifizierungschancengesetz) covers up to 100 percent of course costs plus up to 75 percent salary coverage during training time for SMEs under 10 employees. For 10-249 employees up to 50/50 percent. Important: only training measures of more than 120 hours total qualify (§82 SGB III) — short 1-2-day workshops do not, longer certificate programmes (e.g. IHK / AZAV multi-week or multi-month programmes) do. Application via the German Federal Employment Agency.
Mittelstand Digital Centres: roughly 30 centres across Germany offer free AI consulting, workshops and demonstrators. Direct successor of go-digital — only without invoice. Almost nobody uses the offering fully.
Digitalbonus Bayern between 7,500 and 30,000 EUR for AI/digitalisation investments. Comparable state programmes exist in Baden-Württemberg, NRW, Hesse, Saxony — with different amounts.
Bildungsgutschein (Education Voucher) for affected employees (structural change, elimination of activities by AI): up to 100 percent cost coverage plus salary replacement when the AfA advisor confirms necessity.

If you are a managing director, CFO or HR lead in DACH mid-market in 2026 asking "how do we finance AI buildup?" the typical first answer is "we have to pay this from operating budget." That is wrong in 2026 — and expensive. The funding stack has fundamentally changed: some large programmes are dead, others have been massively expanded, and the most important lever (QCG) is systematically overlooked.

This post delivers the complete overview of active funding programmes 2026, the discontinued ones that are still googled, and a 5-step application workflow you can complete in 4 weeks. Plus: typical mistakes in application.

Who this post is for and who it is not

This post is for managing directors, CFOs, HR leads and IT responsible parties in DACH mid-market (10 to 500 FTE) wanting to start or scale AI initiatives in 2026 and needing to know which funding levers are really available. Concretely: you are either planning a first productive AI initiative (see our 90-day use case matrix), or you have initiatives running and want targeted funding for employee training (see AI literacy mandate post).

Not a fit for large companies above 2,499 employees that no longer fall under most SME funding thresholds. For those the funding stack is structurally different (research tax credit, EU programmes like Horizon Europe).

What expired in 2026 — and why thousands of mid-market firms still search for it

Before we get to active programmes, three programmes that are explicitly no longer applicable. This is relevant because most funding research in mid-market starts with these names and then ends in a dead end.

Digital Jetzt (BMWK): investment grant for SMEs for digitalisation hardware, software and employee qualification. Funding rate up to 50 percent, funding amount up to 50,000 EUR per project (or 100,000 EUR in consortium). Programme ended 31.12.2023. BMWK has not announced a successor. Whoever googles "Digital Jetzt 2026" finds outdated provider pages and no application option.

go-digital (BMWK): consulting grant for SMEs with three modules (digitalised business processes, digital market development, IT security). Funding rate 50 percent, max 16,500 EUR per applicant. Programme ended 31.12.2024. BMWK has since referred to the free Mittelstand Digital Centres as a structural successor.

INVEST – Grant for Venture Capital: proportional reimbursement of investment in young innovative companies — primarily for private investors but partially used by mid-market firms as co-invest structure. Programme was significantly cut in spring 2025 and is practically no longer usable for classical mid-market AI projects.

The consequence: whoever lands in a classical funding research process in 2026 (Google search, funding database keyword) often gets outdated hits. The next 7 programmes are the ones that are really active and that you should focus on.

The 7 active funding levers 2026

1. Qualifizierungschancengesetz (QCG) — the underestimated main lever

What it is: funding instrument of the German Federal Employment Agency. Covers course costs plus partial salary replacement during employee training. Since 2024 explicitly applicable for AI training, expanded again in 2026.

Funding level by company size:

Under 10 employees: up to 100 percent course costs + up to 75 percent salary coverage
10-249 employees: up to 50 percent course costs + up to 50 percent salary coverage
250-2,499 employees: up to 25 percent course costs + 25 percent salary coverage
Above 2,500: no funding via QCG

Prerequisites: certified training via AZAV-approved provider, application BEFORE training start, measure with more than 120 hours total (§82 SGB III — short 1-2-day workshops are explicitly excluded; multi-week certificate programmes or modular courses spanning several months do qualify). Application can be made by management or via external funding advisor.

Where QCG fits — and where not:

✅ Fits: multi-week certificate courses (e.g. IHK / AZAV "AI Manager", "Prompt Engineering Specialist", "AI Champion 6-month programme"), modular programmes that sum to 120+ hours, retraining programmes
❌ Does not fit: short Tier 1 basic training (2-4h), Tier 2 in-depth (8-16h); even Tier 3 specialist (40h) sits below the threshold. For these, Mittelstand Digital Centres (point 2) are the better lever — free and without an hour minimum

Practice example (qualifying scenario): 100-employee mid-market firm trains 5 AI Champions in a 6-month modular certificate programme (160 hours total, AZAV-certified). Gross 25,000-50,000 EUR. At under-10-FTE tier QCG funding leaves net 0 EUR; at 10-249-FTE tier net 12,500-25,000 EUR. Difference: 12,500-25,000 EUR not-claimed funding volume per AI Champion cohort.

Why almost nobody uses it: the funding has existed since 2019 but most managing directors in mid-market do not know it. AI-specific certified courses are documented only since 2024. In Sentient engagement consulting we see that roughly 70 percent of mid-market firms do not know or use the funding.

2. Mittelstand Digital Centres — free AI consulting instead of writing applications

What it is: Germany-wide network of roughly 30 regionally responsible centres, funded by the BMWK as successor structure for go-digital. Offer workshops, demonstrators, individual consulting — all free for SMEs.

What you get for free: AI Readiness Check (KIRC), use case workshops, provider overviews, funding consulting (especially QCG application), industry-specific demonstrators (e.g. AI in manufacturing, AI in trade, AI in logistics).

How you get in: call to your regionally responsible centre (list via mittelstand-digital.de). 30-minute initial conversation clarifies your need. Follow-up appointment typically within 2-4 weeks.

Limitation: Mittelstand Digital Centres are primarily exploration/consulting resource, not implementation partners. For actual AI use case delivery you still need external implementation partners or internal buildup (see make/buy/partner post).

3. Bildungsgutschein (Education Voucher) of the Federal Employment Agency

What it is: when the AfA determines that employees must be retrained due to structural change/AI-related task change, an Education Voucher is issued. Up to 100 percent course costs plus unemployment-benefit-level salary replacement during retraining.

Prerequisites: AfA advisor confirms personal necessity. Training must be certified. Typically applies to employees whose previous activity is substantially changed by AI (e.g. clerical work, quality control, customer service).

Strategic use: if you are planning AI-related restructurings the Education Voucher can be a financial safety net for affected employees — and reduces personnel-adjustment costs in a socially compatible way.

4. Digitalbonus Bayern (and comparable state programmes)

What it is: Bavarian state funding programme, investment grant for digitalisation projects in SMEs. Funding amount 7,500 to 30,000 EUR, depending on application variant (Standard, Plus, Premium).

Comparable state programmes:

Baden-Württemberg: Digitalisierungsprämie Plus (5,000-30,000 EUR)
NRW: Mittelstand Innovativ Digital (consulting + investment grant, up to 30,000 EUR)
Hesse: Digital-Zuschuss Hessen (3,000-25,000 EUR)
Saxony: E-Business funding (5,000-50,000 EUR)
Other federal states: research via the federal funding database

Strategy: state programmes vary significantly in funding amount, application deadlines and prerequisites. Mittelstand Digital Centres consulting (point 2) is the simplest entry into research because they know regional programmes.

5. KfW ERP-Förderkredit Digitalisierung / Innovation

What it is: low-interest KfW loan for digitalisation and innovation projects in SMEs. Not a grant but interest-reduced by typically 1-3 percentage points vs. market conditions. Term 5-20 years. Note: the earlier product "ERP-Digitalisierungs- und Innovationskredit (380)" was migrated into the new ERP-Förderkredit family on 01.07.2025 — that is the current name and application path.

When useful: if you are planning a larger AI initiative (200,000+ EUR investment over 12-24 months) and want to spread liquidity burden over several years. Especially for hardware investments (own GPU servers, on-premise setup) sensible.

When not useful: for pure software licences (ChatGPT, Copilot, Claude Enterprise) loan is over-dimensioned — running costs should come from operating budget.

6. ZIM (Central Innovation Programme Mid-Market)

What it is: federal funding programme for research and development projects in SMEs. Funding rate up to 55 percent for cooperation projects with research institutions, funding amount up to 550,000 EUR per project.

When it fits AI: if your AI project has real R&D character (own model development, industry-specific AI innovations, cooperation with university or Fraunhofer). Not for pure application integration of standard AI tools.

Application complexity: high. Realistically only with external funding consulting or experienced internal innovation manager. Approval rate fluctuates by topic between 30-50 percent.

7. Tax research allowance (Forschungszulage)

What it is: tax credit of 25 percent (35 percent for SMEs) on personnel costs for research and development activities. Also applicable for AI R&D. Substantially expanded by the German Wachstumschancengesetz in 2024: assessment base up to 12 million EUR per year, max funding 3.0 million EUR (SMEs at 35% rate: up to 4.2 million EUR) — the previous cap was 1 million EUR.

When it fits AI: if you employ own AI engineers working on new models, skill library architectures or industry-specific fine-tuning projects. Pure application integration is not funded.

Application: via the Federal Office for Tax Research Allowance (BSFZ). More complex but applicable retroactively (for already-performed R&D activity).

60-minute sparring on your funding strategy →

The 5-step QCG application workflow (in 4 weeks)

The largest funding lever is QCG. Here the concrete workflow for application.

Step 1 (1 week): identify QCG-eligible training tier. QCG only funds measures of >120 hours total. In practice this means: AI Champion programmes (4-6 months, 120-200h), certificate courses to "AI Manager", "Prompt Engineering Specialist" or "AI Architect" (each 120+h modular over multiple months). Sentient Tier 1 (2-4h basic) and Tier 2 (8-16h in-depth) sit below the threshold and are covered separately via Mittelstand Digital Centres or operating budget. Output: list of employees who should run a 120+h programme.

Step 2 (1 week): choose AZAV-certified provider. Provider and measure must be AZAV-certified (German accreditation regulation for employment funding). List via AfA advisors or the KURSNET portal. Bitkom Academy, IHK Academies, larger specialised AI providers are typically certified.

Step 3 (1 week): submit QCG application BEFORE training start. Application form via arbeitsagentur.de. Important fields: training provider, costs, period, affected employees with reason ("qualification for AI use in work environment"), salary coverage application (separate position).

Step 4 (1 week): wait for approval + start training. Approval typically within 2-3 weeks after application. Training may only start AFTER written approval — otherwise funding is lost.

Step 5 (ongoing): documentation for compliance file. Training participation confirmation, course content, possibly test result. Simultaneously serves AI literacy compliance from August 2026.

The typical three mistakes in funding applications

Mistake 1: submit application AFTER training start. QCG (and also Digitalbonus, ZIM, KfW loan) require application BEFORE measure start. Whoever has already booked and started training gets no funding anymore. Common mistake with "spontaneous" training bookings.

Mistake 2: measure under 120 hours or non-AZAV-certified provider. QCG only funds AZAV-certified measures with more than 120 hours total scope. Short in-house workshops, 1-2-day trainer bookings or boutique consultants without AZAV approval get nothing. Check measure duration AND certification status before booking.

Mistake 3: overlook Mittelstand Digital Centres consulting. The centres are free and know regional funding programmes better than any funding consultant. Whoever goes into own research without this consulting conversation typically overlooks 30-50 percent of relevant programmes.

Frequently asked questions

We have under 10 employees, is funding worth it at all? Yes — at under 10 employees you get the maximum QCG funding rate of up to 100 percent course costs plus 75 percent salary coverage — provided the measure clears the >120-hour threshold. For a 5-employee team in a 6-month AI Champion programme that is effectively free qualification. For short 1-day workshops QCG does not apply — there the Mittelstand Digital Centres are the better lever.

What if our employees are not unemployed, do we still have a claim? QCG is explicitly designed for employed personnel without unemployment. Bildungsgutschein is for job seekers or employees acutely affected by structural change. Both programmes usable in parallel.

Can we apply for funding retroactively? No for QCG, Digitalbonus, ZIM, KfW. Yes for tax research allowance (retroactive for 4 years). Plan: ALWAYS submit applications before measures, otherwise funding lapses.

How long realistically does application take? QCG: 1-2 hours internal preparation plus 1-2 weeks approval period. Digitalbonus Bayern: 4-8 hours application creation plus 4-8 weeks approval. ZIM: 40-80 hours plus 3-6 months approval. Scaling from simple to complex.

Is an external funding consultant worth it? For QCG no — the application is manageable, HR lead handles it. For ZIM, larger investment projects or state programmes with complex requirements often sensible because success rate doubles-triples. Consulting fees usually 8-15 percent of approved sum.

What if our headquarters is in Austria or Switzerland? This post refers to German funding programmes. Austria has similar programmes via FFG and Wirtschaftskammer (e.g. digitalisation funding, AI-specific programmes). Switzerland has via Innosuisse R&D funding. Both are basically active in 2026 — the German programmes of this post are not accessible to you however.

Which AI stage are we? The 15-minute maturity check →

AI Funding for DACH Mid-Market 2026: 7 Active Programmes Plus the Discontinued Ones Everyone Still Googles