CIR and CII approval: how the French State funds up to 30% of your projects

What the approval changes for you, not for us

CIR-CII approval is granted by the French ministry responsible for research. It certifies that a provider has the skills and resources to carry out research or innovation work on behalf of a third party. The key point is this: it gives us no tax advantage; it gives one to our clients. Without approval, the sums a company pays to a private provider simply cannot be included in the base of its own tax credit — the expense is real, the invoice exists, but it does not count. With approval, the same invoice becomes an eligible expense on your side.

In practical terms, a development project carried out with us does not cost you what it costs: part of it comes back to you the following year, as a claim on the State or as a refund. It is the only innovation support scheme that requires no grant application, no selection committee and no review period: it is declared with your tax return. Approval is generally granted for three years, and the ministry publishes the list of approved organisations every year.

Two schemes, two rates, two audiences

People often talk about the “CIR” as if it were a single thing. In reality these are two distinct mechanisms, which do not target the same work.

The Crédit d’Impôt Recherche (CIR, research tax credit) covers research and development in the strict sense. Its rate is 30% of eligible expenditure up to €100 million of base, and 5% above that. It is open to all companies, whatever their size, provided the work genuinely qualifies as R&D.

The Crédit d’Impôt Innovation (CII, innovation tax credit) is reserved for SMEs as defined by the EU — fewer than 250 employees, and turnover below €50 million or a balance sheet below €43 million. It covers the design of prototypes and pilot installations for new products. Its rate deserves particular attention: it has been 20% in mainland France since 1 January 2025, compared with 30% in the two previous years. Much online content still shows the old rate. The expenditure ceiling is €400,000 per year, giving a maximum credit of €80,000, and the scheme has been extended until 31 December 2027.

That is why we write “up to 30%” and not “30%”: the rate depends on which scheme your work falls under, and it is the CIR that carries the higher rate.

Sources: innovation tax credit on Service-Public, and article 244 quater B of the French General Tax Code (CGI) on Légifrance.

What is eligible — and what never is

This is the part where we regularly disappoint, and we would rather do so before signing than after a tax audit.

Never eligible: maintenance of your IT estate, managed services, troubleshooting, workstation set-up, network cabling, installing a firewall or cameras, migrating an email system, configuring off-the-shelf software. These are useful services, often urgent, but they are not research work. An invoice issued by an approved provider does not become eligible for that reason: what counts is the nature of the work, not the status of whoever invoices it.

Software development is not eligible by default either. Writing a business application, even a bespoke one, even a complex one, is routine development: the techniques are known, the outcome is certain from the outset, and only the amount of work is at stake. To qualify for the CIR, a project must come up against a technical barrier that the state of the art cannot overcome, and overcome it through an experimental approach — with trials, failures and measurements. The question to ask is not “is it difficult?” but “is there a known solution, and if so, why is it unsuitable?”.

There remain real cases, which we do encounter: document processing that must reach a level of reliability that available models do not achieve on your business corpus; a response-time or data-volume constraint that no documented architecture can meet; a product prototype you want to bring to market and that nobody yet knows will work. The first and third cases often fall under the CII rather than the CIR. This is precisely the sorting we do at the scoping stage.

What it really amounts to

Take a document assistant project for a professional firm, with a €60,000 budget, of which a genuinely experimental part is assessed at €25,000 — the rest being integration, data migration and training, which are not eligible. If that €25,000 falls under the CII, the credit is €5,000; if it falls under the CIR, it is €7,500. The project does not become free, but its net cost falls by 8 to 12%.

Two ceilings frame this mechanism, and you need to know them before building a funding plan. First, subcontracting expenditure is only taken into account up to three times the amount of the company’s other research expenditure: a company with no in-house R&D activity therefore cannot fund an entirely outsourced project. Second, a monetary ceiling applies — €10 million per year where there is no dependency relationship between the client and the provider, €2 million where there is. The first ceiling is the one that catches SMEs out; the second in practice only concerns large accounts.

Key points

• The approval benefits the client, not the provider: it makes the invoice eligible on your side.
• 30% is the CIR rate; the CII has been at 20% since 1 January 2025, capped at €400,000 of expenditure per year.
• Maintenance, managed services, hardware and networking are never eligible, whoever the provider.
• Development, even bespoke, is only eligible if it resolves a documented technical uncertainty.
• Subcontracting is capped at three times your other R&D expenditure: a 100% outsourced project does not qualify.

How we work

We do not sell tax structuring and we do not file anything on your behalf — that is the job of your chartered accountant or a specialist adviser, with whom we are happy to work. Our role is limited to three things, and they are decisive in the event of an audit.

We tell you frankly, at the scoping stage, what does and does not count as R&D in your project, and we cost the two parts separately. We invoice clearly: an “R&D work” line separate from the integration and training lines, because a single €60,000 invoice headed “IT services” cannot be defended before an inspector. We document the work as it progresses: the technical barrier targeted, the state of the art reviewed, the hypotheses tested, the results obtained, including failures. This documentation is your evidence, and it cannot be reconstructed two years later.

The costly mistakes

Three come up again and again. Declaring first, documenting later: a file reconstructed after the fact shows, and the tax authorities challenge the credit with late-payment interest. Confusing novelty for the company with novelty for the state of the art: a technology being new to your teams does not make it new in the CIR sense — this is the criterion that sinks the most claims. Letting the tax credit drive the investment decision: if a project only makes sense because of the credit, its practical value is not enough, and you will end up with a tool nobody really wanted. The tax credit improves the return on a justified project; it does not justify a project.

If you have a project in mind and do not know which side of the line it falls on, tell us: the sorting takes one conversation, and it is better to have it before the quote is drafted.