AI Training

FNE-Formation AI: what actually funds your SME AI training in 2026

For years, FNE-Formation was the most generous way to fund AI training in France, covering up to 70 % of costs. It can no longer be used for new training in 2026. Two other levers now take over, and one still funds up to 50 % of your AI training.

RRomain BellaicheJune 5, 20268 min read

Read the full guide

AI Training for Business: The Complete 2026 Guide (Funding, Certification, EU AI Act)

FNE-Formation AI: what actually funds your SME AI training in 2026

For years, FNE-Formation was the most generous way to fund AI training in France, covering up to 70 % of costs. It can no longer be used for new training in 2026. Two other levers now take over, and one of them still funds up to 50 % of your AI training. Here is which ones, and how to build an application that gets approved. This article complements our complete guide to corporate AI training.

FNE-Formation is no longer open for new AI training in 2026

In practice, you can no longer submit a new FNE-Formation request to fund AI training. The scheme, run by the Ministry of Labour as part of support for economic change, worked through agreements signed with the DREETS or via the OPCOs. According to information published on travail-emploi.gouv.fr, funding agreements had to be signed by 31 December 2024, and training had to end by 31 December 2025 at the latest.

No new FNE-Formation budget has been reopened for 2026. [To be checked with your OPCO advisor before any submission, as terms may change during the year.]

Many articles online still present FNE as available with 70 % rates. That information is outdated. To fund an AI upskilling project today, start from the schemes below instead.

What FNE-Formation funded, and why it still matters

FNE-Formation historically covered up to 70 % of training costs for companies under 300 employees, and 50 % above that. It targeted transitions, in particular the digital transition, which includes artificial intelligence. Indirect costs and employee wages were not covered.

Knowing this logic still helps in 2026, because the closest current scheme, FSE+, works on the same principle: public co-funding of part of the training costs, provided you go through a certified provider and apply before the start date.

FSE+, the closest replacement for AI training

The European Social Fund Plus, or FSE+, currently funds up to 50 % of the training costs of a digital-transition course, AI included. You access it through your OPCO, which acts as the programme operator. Eligible topics cover the digital transition broadly: software onboarding, technical training, generative AI.

Two points to watch. First, the current FSE+ programme is extended until 30 June 2026, with submission deadlines that vary by OPCO. As an indication, some operators announce deadlines spread between March and June 2026. [To be checked with your OPCO: the exact deadline and remaining budget determine approval.] Second, as with FNE, wages and indirect costs are not covered: only training costs are.

If you aim for FSE+, the timeline is tight. An application filed after the training starts, or after your OPCO deadline, is rejected outright.

The skills development plan, the durable route

The most stable funding from year to year remains the skills development plan, run by your OPCO. Unlike FNE or FSE+, it does not depend on a time-limited exceptional budget. For companies under 50 employees, coverage of training costs is often high, sometimes 100 %, depending on the sector and available budget.

The full process is detailed in our article on funding AI training through your OPCO. The principle: identify your OPCO via your IDCC code, choose a Qualiopi-certified provider, then file the funding request before the first day of training. Without the provider's Qualiopi certification, none of these schemes can be used. We explain why in our article on Qualiopi AI training.

Comparison of schemes available in 2026

SchemeStatus in 2026Rate on training costsFor whom
FNE-FormationClosed to new requestsHistorically up to 70 %No longer available
FSE+Open, limited window until 30 June 2026Up to 50 %SMEs, via OPCO, apply before start
Skills development planOpen continuouslyVariable, often high for under 50 employeesAny company, via OPCO
CPFOpenEmployee's individual creditAt the employee's initiative

The CPF remains an individual option. If an employee wants to fund an AI certification on their own account, the 2026 rules are covered in our article on the CPF for AI training.

How to build an AI funding application that gets approved

An approved application is prepared before the first day of training, never after. Here are the steps that make the difference:

  1. Identify your OPCO. Find your IDCC code on a payslip, then match it against the correspondence table published by the Ministry of Labour.
  2. Choose a Qualiopi-certified provider and ask for a quote with a detailed programme, learning objectives and assessment methods.
  3. Check the applicable scheme and its deadline. For FSE+, your OPCO deadline takes priority over the programme end date.
  4. File the funding request before the start, attaching the quote and programme.
  5. Keep the completion certificate and attendance sheets: they condition the payment.

To frame the overall budget and compare real costs, our article on the price of corporate AI training gives the 2026 ballpark figures.

The mistakes that get funding refused

The most common cause of refusal is late filing. A request sent after the first day of training is systematically rejected, whatever the scheme. Next come using a provider that is not Qualiopi-certified, a programme too vague to attach to an eligible topic, and missing the FSE+ specific deadline.

One last confusion comes up often: believing that funding covers employees' salaries during training. It does not. These schemes fund training costs, not time spent. That is one more reason to size the upskilling effort well so it stays profitable even with working time included.

Frequently asked questions

Does FNE-Formation still exist in 2026? FNE-Formation no longer accepts new requests. The last agreements had to be signed before the end of 2024 and the training completed before the end of 2025, according to travail-emploi.gouv.fr. No new budget has been reopened for 2026 to date.

What replaces FNE-Formation for AI? FSE+ is the closest, with up to 50 % coverage of training costs, in a limited window. Your OPCO skills development plan is the most stable option over time.

What funding rate applies to AI training in 2026? Up to 50 % via FSE+ on training costs, and often more via the skills development plan for companies under 50 employees. The exact rate depends on your OPCO and sector.

Do you need a Qualiopi provider to get funding? Yes. The provider's Qualiopi certification is a condition for any public or pooled funding, FSE+ as well as the skills development plan.

Can you combine several funding sources? Combining is regulated, and it is rarely possible to fund the same cost twice. Your OPCO advisor checks compatibility case by case. [To be checked before submission.]

How long before the training should you apply? As early as possible, and always before the first day. Allow several weeks of processing depending on the OPCO. For FSE+, the OPCO deadline sets the cutoff.

How GrowthPerf supports your funding requests

GrowthPerf is a Qualiopi-certified training provider specialised in AI, no-code and automation for SMEs and non-profits. We build the application with you: identifying the right scheme, quotes and programmes that match OPCO expectations, and on-time submission. You keep control of the learning content, we secure the administrative side.

Want to know which funding applies to your situation before the FSE+ window closes? Ask for help building your application. We review your OPCO, your deadlines and the achievable coverage rate together.

R
Romain Bellaiche

AI & No-Code Expert

Assess your AI maturity in 3 minutes

Answer a few questions and get a personalized assessment with recommendations tailored to your industry.

Assess my AI maturity