Late 2020, Alexandre Yazdi, founder and CEO of Voodoo spoke at the Elysée Palace during the La French Tech Next40 event and stood to say that “it would be necessary to position the unicorn as an intermediate stage before something more ambitious, that is to say to create from France world giants of technology” adding besides that France has the necessary resources to achieve this goal.
Alexandre Yazdi is right when it comes to creating a tech giant. Indeed, companies purely focused on Artificial Intelligence have a significant advantage : by nature, AI is scalable everywhere and its market is global, wherever the company is based.
According to Kai-Fu Lee in his book AI Superpowers, to create AI superpowers you need to meet certain criteria: abundant data; be fierce entrepreneurs; trained scientists and a favourable political environment.
So where are we in France and what are our resources and assets to create AI giants? How can France leverage its traditions and specificities to differentiate its position on the AI global scene?
Success factors at the scale of the economy
Data
First, access to abundant data is one of the key factors of success for developing an efficient AI solution. This is the starting point of any AI strategy as data is in the hands of only a few actors. In his report, Cedric Villani wanted to reflect on the production and collaboration of data by constituting a “data commons”, so that economic players, public authorities, and research can benefit from it and rebalance the balance of power with the major players.
Due to intense GDPR measures protecting citizen’s personal data, France and Europe are often seen as side-lined when it comes to data collection. Personal data concerning consumer habits, travel or tastes are thereby difficult to obtain (and fortunately), and therefore less accessible in France and Europe, except for GAFAMs. Massive data collection being the very essence of any AI business, should this mean that France will remain an outsider of the AI global market?
Absolutely not. France can benefit from a wide ecosystem of well-digitized companies that have stored massive amounts of data in a very structured way (which is essential for any machine learning application) in high level fields such as industry (energy, nuclear, aerospatial), finance or insurance. Therefore, French AI players shouldn’t try to imitate their American and Chinese competitors, but should build their own identity, relying on these well qualified, top of the art databases to train their algorithm.
The French entrepreneurship ecosystem
The French have a thirst for entrepreneurship: nearly one out of three French people would like to start their own company and the number of companies created since 2009 is fast growing. This is partly due to the “auto-entrepreneur” status created in 2008 but not only.
In 2020, France set a record with 850 000 new businesses created (4% more than 2019’s record), 220 000 of which are traditional companies (counting more than one person). In the past few years, the company creation rate has been rising and it is especially outstanding regarding the economical context induced by Covid-19 in 2020.
Thus this rising craze for company creation in France is not only due to this French thirst for entrepreneurship, but also to the favourable ecosystem built in France in the past few years: according to Global Entrepreneurship Development Insitute (GEDI), France had the 10th healthiest entrepreneurial ecosystem worldwide in 2018 (while Germany has the 15th one).
The French government is indeed fostering innovation and entrepreneurship through initiatives, subsidies, investment, and legal frameworks. La French Tech Initiative was launched by the French Government at the end of 2013 to build a the French Tech solid group, highlight what already exists, and create a snowball effect.
More recent incentives have also been launched, like the DeepTech Plan 2019 (800M€) and the Plan de Relance 2020–2022, in which 3,7B€ will be allocated to fostering the development of start-ups and of the French technological sovereignty.
Moreover, Bpifrance Le Hub, is supporting entrepreneurs and investing in startups,. Its goal is to support French innovation from school labs to Nasdaq. It also supports companies with public loans with no interest rates in difficult times like 2020. Finally, public policies like French Tech Visa or the Loi Pacte are supporting talents’ immigration and simplifying the legal framework to favour startups development.
Tech & AI ecosystem
The French tech ecosystem is very active and attractive. Just look at the latest fundraising events, the VC activity (most active european country since 2017) and the success stories of the last ten years in France.
We hear much more about French startups doing large fundraising and having an ambitious international roadmap with plans to become AI giants. The French see their American and Chinese counterparts as well as their predecessors in the French Tech succeed and now they just happen to want that, which is a paramount factor for the creation of tech giants.
Corporations from the CAC40 have almost all launched corporate VCs, incubators, accelerators and are mostly trying to step in AI. In Paris, incubators like Agoranov are well established and have welcomed a few unicorns (Dataiku, Criteo France, doctolib) and are now welcoming the French gratin in AI.
At Monk ai, we think that smart data collection is key to our development and of the further development of our machinelearning solution. We therefore had the opportunity to experience the great qualifications of French companies in terms of data collection and structuration, in particular with Drivy (Getaround) in the early stages of Monk development. They provided us with a massive and qualified amount of car damages data that enabled us to practice our AI solution, making us one of the most accurate providers in this field.
Research
France has a long lasting tradition of excellent and important public research (2,24% of the GDP allocated to R&D, 5th rank worldwide just behind the US) but private research is mainly less powerful as in the US or the UK.
In the context of artificial intelligence, research is predominantly private. To make it work with public research, one of the pillars of the French IA strategy is to bring private and public IA research together through the “3IA” (Instituts Interdisciplinaires d’Intelligence Artificielle), which are interdisciplinary centres meant to be the flagship of the IA strategy.
In parallel, many incentives have been created to foster private research, among which the CIR, which allows companies to be exempted on taxes on their R&D salaries expenses and JEI for fiscal advantages. (It has been estimated that each euro invested by the state in the CIR will result in a GDP creation of 2,3 to 4,5€ in the long run).
Education
French engineering programs are well-known for their rigorous curricula aimed at positioning graduates for successful careers following graduation. They also integrate business training, foreign language study, and communication skills for delivering creative solutions to some of today’s most complex challenges. French engineering schools are especially attractive despite their entry selection based on the mathematical level and concentrates most of the country’s elites. If there is excellent research in France there are difficulties in transforming scientific advances into economic applications. The main reason is that R&D is not entrepreneurship. French engineers tend to overthink their problems instead of getting their hands dirty.
Recently, the French education system has started training its engineers to the digitaleconomy and business matters and its business graduates to data related positions, coding and AI concepts. French best business schools (HEC Paris, Essec, ESCP) are offering alongside top engineering schools (X, Centrale Supélec) double Degrees and are training new generations of polyvalent graduates who can endorse the difficulty of being a tech entrepreneur these days.
There is no doubt that France has the fitting education ecosystem for creating the AI leaders of tomorrow, what can already be noticed by seeing the important number of French executive officers in the best Tech and AI companies in the world (Yann LeCun won the Turing Award 2019, Luc Julia was one of the founders of Siri). The stake today is to encourage those talents to stay in France and create business here.
Conclusion
Was Alexandre Yazdi right to mention a more ambitious goal than just reaching the unicorn stage ? Absolutely. The La French Tech scene has everything to start creating tech giants. Data, entrepreneurs, engineers, research and a dynamic political framework and domestic market to grow and reach that stage.
France needs to gain confidence in its assets, resources and differences and exploit them at best to differentiate its position on the AI competition with China and the US (notably by exploiting the different types of data that it has gathered and leverage it).
Now that all of the ingredients are available and gathered together, the hardest part is still in front of us. It is most likely that in a near future, fierce entrepreneurs will gather top engineers and use available data and government resources to create an AI leader that will then become a giant. At that time we’ll look back at 2020 saying it was obvious since France had all of the factors in its favor to do that. Now let’s do it.
Aboubakr Laraki
