He repeats the envy that he has one of the best positions in the Republic. Bruno Bonnell is responsible for a maximum of 54 billion euros reserved for innovation. The secretary general dedicated to the investment in charge of the plan “France 2030” has also crossed the legislatures and governments since January 2022 without trembling. But everything is not so rosy. The budget deficit is currently obliging him to play the balancingists in his accounts. There are finally many criticisms on his future choices: will they be effective? Does it sprinkle the money too much available? His results perhaps only have an effect in several years. And you must constantly adjust the sights. In artificial intelligence, the “Deepseek moment” – named after this Chinese start -up which is currently upsetting the market – comes to question the investment policy. Bruno Bonnell must set the course. Interview.
https://www.youtube.com/watch?
L’Express: in the race for artificial intelligence, the case of The Chinese start-up Deepseek seems to indicate that the money may be less decisive than we imagined to be competitive. What lessons do you learn from it?
Bruno Bonnell: A teaching is that the strategic response for France and Europe is probably to accelerate on open source. This open base is a good strategy of response to American paid licenses, in order to counter their domination. Modern digital is also based on a multitude of open source software, such as Linux. It must be, as such, that we support Mistral, it is not even a debate. The state, all private capital, all companies must be behind. Economically, it’s viable. Because we can then build a lot of applications from this base. We can thus try to obtain in Europe the best champions in medical, cultural AI … For this, we will in France 2030 double the put in AI, from two to four billion investments.
But the real subject is not only economical, it is cultural: we cannot lose our identity through tools that reflect Chinese or American values and culture. Open source can provide the necessary transparency to an ethical and confidence AI. AI is a determining element of the transformation of societies. It must be supported on human values, not just competitiveness. Our strategy must take it into account.
Especially at a time when the pressure exerted by the United States on Europe is intensifying, with the Trump-Musk duo…
There is a choice to make, either accept the situation and “take advantage” of a technology completely imported from the United States or even China, from mobile phones to autonomous cars. Be “vassalized”, even “techno-fascized”. Or realize that AI is a societal transformation lever and define our own strategy according to our values. We have already done so in many other areas, for example firearms.
However, it is necessary to be able to count on reliable technologies. Recently, An open source French chatbot called “Lucie” Has talked about him a lot for his many mistakes. The project, partly funded by France 2030 funds, was disconnected in just three days. Do you regret having supported it?
No, I don’t regret it. France 2030 is made to test and select the best technologies. Lucie IA was an open source model of language model. It was not yet mature, but that is not a reason to abandon. Ariane’s first trials also failed … And yet France is very efficient in aeronautics. Failure is part of the innovation process. And it must be remembered, in France 2030, less than 5 % of projects fail. The important thing is not to consider them as death sentences, but as learning and improvement opportunities.
The Senate recently amputated 535 million euros from the France 2030 budget for the year 2025. Do you fear, in the current budgetary climate, that the plan be planed?
First, I want to specify it from the outset: this does not impact the 54 billion euros of the France 2030 plan. The 535 million receive payment credits, that is to say it is the checks that we do during the year for supported projects. So we are talking about a gap and not a planing. The investment, so far from a total of 38 billion euros, will just be smoothed. This shifts future commitments, with reasonable proportions all the same compared to the overall envelope which remains, around 15 billion euros.
The France 2030 Investment Plan is again criticized for its dusting in a recent report by the Court of Auditors. Does this annoy you?
Yes, because that does not reflect reality. France 2030 is based on 10 clear objectives, broken down into interdependent technological bricks. For example, building a low carbon plane by 2030 requires advances in materials, fuels and motorization. We work on all these dimensions in a coordinated manner, and it is not sprinkling. The state takes risks by betting on innovation. It will take time, but that’s how we build the future. Without daring decisions in the 1970s, we would have neither TGV nor nuclear program.
And it should be remembered, the State replaces itself in an exceptional way in France for the failure of private capital for breaking innovations. When I read that Donald Trump places $ 500 billion in AI, the US state does not actually put a penny. These are companies like SoftBank or Oracle that decide to collectively put 500 billion. We do not have the same land of private investment in France or even in Europe.
What assessment do you draw from France 2030 to mid-term?
For the time being, 80,000 industrial jobs have been created since 2017. At 2030, anticipations remain around 300,000. Regarding the envelope, there are therefore about 15 billion to engage, on which it is advisable to wonder. Compared to the trajectories that we already have on the different objectives, should we add fuel somewhere? Should we move means on others? This is the discussion that we have continuously. But there are two certainties. On the one hand, we should not lower the level of support. In particular, it is necessary to maintain the priority of training and research. We have already opened 240,000 training places in hydrogen, nuclear, new agriculture, health. And of course, put the package on AI. As I said, we will double the efforts. It is not the catch -up that interests me, but the projection in the innovation of tomorrow.
.