Technology is stuff that doesn't work yet

France is no tax heaven, but for technology companies it may well be Research & Development heaven: there is a lot of talent coming from the Grandes Ecoles and the country offers two schemes aimed at helping fund R&D that are arguably some of the most interesting in the world, says Rémi Amengual of Lowendalmasaï (computer science and mathematics department). These are the Jeune Entreprise Innovante (young innovative company - JEI) status, which concerns startups exclusively, and the Crédit Impôt Recherche (research tax credit - CIR) which may be one of the reasons Google and Microsoft have a strong R&D presence in France. See section 1 of the YIC Handbook for more information in English.

Public funding for innovative tech startups

The JEI status entitles you to tax breaks/exonerations if you are a startup that creates innovative technology and spends at least 15% of its resources on R&D. You will need to justify at some point that what you are doing (or what you are planning on doing) is valid R&D — more on that below. The best way to proceed is to ask for a rescript, i.e. an assessment of your eligibility to the scheme. The ideal time to file your request is in september so you get a reply by december of year Y-1. If your project is approved, you will also benefit from the CIR which will reimburse a good chunk of your R&D expenses at the end of year Y — but you still need to have enough cash flow to cover for all your charges in the meantime. Actually, the CIR takes the form of tax breaks for larger companies, and credit for JEIs who don’t pay taxes (or small amounts).


In the first couple of years you can hope to get:

  • virtually no taxes
  • 60% off salary of (software) engineers, multiplied by the proportion of time they spend on R&D
  • 120% off salary of recent PhD graduates (yes, you get more money than you spent!)
  • 80% off contracting with European universities, organisations or freelances that have been approved by the French ministry of research.

As you would have guessed, these percentages decrease as your company gets older — don’t tell me you’re thinking more than 2 years ahead ;) . Besides, the amount of money you may get from public grants will be subtracted from the amount of CIR money you can claim — which makes sense for the State.

The calculations are actually quite complex if you want to be accurate, so you may want to get in touch with specialists when writing your business plan. They will also be able to advise you on the submission of your rescript request, and to help you in the process of declaring/justifying expenses and time spent on R&D activities. For this service you will be charged a fixed fee or success fees (i.e. a percentage of the money the State gives back to you). The latter option seems to be more popular for startups.

R&D vs innovation

It is important to be clear about what R&D is when describing your R&D project in order to get scientific validation from the DRRT on your eligibility to the JEI and CIR. The guideline is that if you already have a pretty accurate idea of how to solve the problem you choose to address, then it is innovation you are doing; otherwise, if the state of the art does not allow you to reach your final objective, if you can explain why things are difficult (on a technical level), identify uncertainties, and plan experiments in order to remove these uncertainties, then it is R&D you are doing. Innovation is the phase that follows and there are dedicated schemes from Oséo for funding it.

Lab Tour

R&D activities consist of fundamental and applied research, as well as experimental development (trials, simulations, prototypes) performed in order to create new products, processes, services, or to substantially improve existing ones. Innovation (in the general sense) is often incremental, and here the increments must have significant technicality. The outcome of research is a contribution to the available scientific and technical knowledge — but this contribution doesn’t necessarily have to be public as in academic research! The process sort-of goes like this: set ambitious objective, review current knowledge, identify difficulties/uncertainties in bridging gap with objective, develop ideas to address/remove them, try them out, interpret results, iterate.

For the CIR, your rescript request will start with a description of your project, its context, and a proposed timeline. The scientific points to be addressed are: [partly translated from French]

  • Background review of the state of the art. You should be referencing recent conference and journal papers. (This will help you make sure you’re not going to waste time reinventing the wheel. Google scholar and Mendeley are your friends and just need be fed with a few keywords.)
  • Objectives set, performances to be attained and constraints.
  • Scientific and technical constraints, obstacles and problems to solve.
  • Any related work you have carried out or results you have obtained so far.

If you want to add substance to your R&D project, you should be adding constraints, technical difficulties, uncertainties… so it becomes clear that you are being pushed to be scientifically and technologically ambitious!

CIR-eligible activities: design, build, run

Any time spent on R&D activities by employees or contractors with a tech CV (“Bac +2” at least, i.e. 2 years at university) can be claimed in the CIR, as long as you’re able to report on progress and substantiate your claim. In the case of software engineering, activities that are eligible are:

  • designing and implementing algorithms
  • writing, testing, optimising, standardising code/processes
  • technology watch (e.g. going to conferences)

Activities that are not eligible:

  • deploying software into production
  • maintenance
  • training
  • benchmarking
  • performing statistical tests
  • designing/building interfaces
  • porting to different platforms
  • debugging

In addition to time spent, you can also claim expenses related to:

  • equipment you need in order to carry out R&D activities
  • the protection of your Intellectual Property
  • running costs (e.g. renting compute instances in the cloud)

For more on the CIR, check out this guide if you read French.


Startups in France can benefit from public, non dilutive funding through the JEI and the CIR, and it’s interesting to see how the latter favours hires of recent PhD graduates — my case! France seems to be very keen on pushing technological innovation in the form of ambitious R&D projects. The country gives means to partly support these ambitions, but startupers are still left with a cash-flow problem. Public organisations such as Oséo can help you get interest-free loans of 50% of the amount of your planned expenses, but you’ll have to manage the rest.