French President Emmanuel Macron was beaming with joy during a press conference with Microsoft president Brad Smith held on the sidelines of a special investment summit in Paris earlier in May.

The US software giant had just announced additional investments of €4 billion ($4.3 billion) in data centers and the artificial intelligence (AI) sector in France by 2027.

"[Microsoft's new] data centre will be one of Europe's biggest and help us be one of the data storage and AI leaders," Macron claimed.

Last summer, France published a national AI strategy with €500 million to be invested in the creation of AI clusters by 2030. A few months later, in December, Paris startup Mistral AI joined the league of AI champions by becoming a  so-called unicorn in the sector — companies valued at more than $1 billion.

Europe lagging behind US and China heavyweights

Noah Greene from the AI Safety and Stability Project at Washington-based think tank Center for a New American Security (CNAS) says the French government "flipped a switch" when it decided to become an AI champion. But making that ambition a reality might be an uphill struggle, he told DW.

With the United States being the clear AI market leader, and China coming second ahead of the UK, the backlog of EU leaders France and Germany was not only down to technological factors, he said. "The US has been at the top of the game for so long that investors prefer to put their money here, as they know the institutional talent and infrastructure already exist."

France meanwhile has a "very complex labor code and large US tech firms like Google have struggled at times to get past these laws," he added.

France 'has excellent AI researchers'

But Veronique Ventos, co-founder of Paris-based startup NukkAI, says she never considered French labor laws a hurdle.

"We always knew we'd set up our company in France with its excellent researchers and numerous support programs for startups," Ventos, who used to work as an AI researcher at Paris-Saclay University, told DW.

Ventos claims the company's AI is different from others because "humans are fully integrated."

"They can observe its processes and are told why the AI is making certain recommendations and decisions," she explained, adding that the technology used considerably less data than other AI algorithms and was thus more energy-efficient.

At the moment, NukkAI has half a dozen clients, including French aerospace group Thales and NATO, which use the technology to plan their logistics.

The startup also closely cooperates with French universities, which gives NukkAI access to France's Jean Zay supercomputer. Based in Saclay in the southwestern suburbs of Paris, Jean Zay is one of Europe's most powerful computers with a capacity of 36.85 petaflops — equivalent to several quadrillions of operations per second.

The race for computing power

Christine Dugoin, associate professor at Paris' Pantheon-Sorbonne University's Artificial Intelligence Observatory, thinks France, and all of Europe, needs more and bigger supercomputers. "That's the only way we'll be able to compete in the area of AI," she told DW.

Additional supercomputers will be inaugurated this year and next in Jülich, Germany, and in France's Essonne department near Paris. The machines will be the first in Europe to exceed a capacity of 1 exaflop per second, which is 1 quintillion operations.

"But we'll still be lagging behind the US and China, which is now claiming to have overtaken the US with its new machine, Tianhe 3, supposedly surpassing a capacity of 2 exaflops, which would make it the world's fastest," Dugoin said.

She thinks a Europe-wide AI approach is needed — and not only to deal with global competition: "Since Russia started invading Ukraine in 2022, it's been deploying an AI-based disinformation campaign against Europe. The only way to fight back is by joining forces."

Joining AI forces in Europe

The German defense company Helsing wants to do just that. Also a unicorn by market value, the Munich-based AI firm was founded in 2021, and also has offices in the UK and France.

Helsing vice-president AI Antoine Bordes says Russia's war in Ukraine war has shown that Western democracies face "an existential risk and need to strive for a form of common technological and defense sovereignty with AI at its center."

Helsing analyzes data from satellites or radars in real time and provides them to troops on the ground, in the air and at sea. It is also working with the Ukrainian military. Bordes told DW that it would require "a Europe-wide AI investment plan, also with regards to our computing capacity" for Europe to catch up with the US and China.

Philippe Aghion, a professor at the College de France, INSEAD and theLondon School of Economics, agrees. In March, he co-authored a French government-commissioned report that called for more state investment in AI.

"The AI sector could boost French GDP by 0.8% per year over the coming 10 years," he told DW, but that potential could only unfold if the government deploys "a proper industry policy and invests at least €25 billion" in the sector.

Greene, however, isn't so sure that Europe will be able to use its potential. "The US has been implementing a 'laissez-faire' policy and putting in place as few hurdles as possible. The EU, by contrast, is aiming to become one of the leaders in the regulation of the technology to protect fundamental rights."

And indeed, in March the European Parliament adopted a so-called AI Act that bans the development of certain AI technologies deemed too dangerous. Greene argues that only AI technology leaders will be able to "control the keys to the castle and decide how autocracies will use these products."

Veronique Ventos from startup NukkAI thinks that we shouldn't even try compete with AI giants the likes of Google "in areas where they are clearly in another league" such as data storage. "We should focus on our own strengths, such as, in France, technologies that combine AI with robotics."

Edited by: Uwe Hessler