AI and the Paris–London rivalry: a tale of two innovation hubs edit

12 June 2026

Artificial intelligence (AI) has become one of the main drivers of economic competitiveness, technological sovereignty, and investment attractiveness on a global scale. While the United States and China remain well ahead in the race, Europe is still striving to establish itself as a major AI power. Three European countries, however, stand out for their ambitions and performance. The United Kingdom maintains a significant lead in terms of its entrepreneurial ecosystem and startup funding, while Ireland and France are now establishing themselves as Europe’s leading hubs for investment in computing infrastructure and data centers dedicated to AI. Ireland got off to a flying start but is now running up against constraints in electricity supply: data centers consume 22% of the electricity produced in the country, posing a high risk to grid stability on an island where renewables account for nearly 40% of the electricity mix. We shall therefore focus primarily here on the British and French strategies, which do not face the same limitations.

The United Kingdom, European Leader in AI Innovation Funding

For several years now, London has established itself as the leading European hub for tech startups, benefiting from a mature ecosystem that brings together world-class universities, venture capital funds, major tech groups, and international talent. This leadership position naturally extends to artificial intelligence.

By 2024, London-based AI startups had already raised approximately $3.5 billion, compared to $2.4 billion for those in the Paris region. Nationally, venture capital investments in AI in the United Kingdom reached nearly $13.8 billion in 2025 according to the OECD, representing approximately 5% of global investments in this field. This performance places the United Kingdom well ahead of other European countries. For that same year, 2025, venture capital investments in AI across the EU-27 totaled $15.8 billion, representing 6% of the global total.

Much attention was paid that year to the €1.7 billion raised by Mistral AI, which was undoubtedly encouraging news. More recently, in March 2026, Yann Le Cun’s Paris-based startup AMI Labs raised €1 billion. But behind these successes, in entrepreneurial terms France is not as dynamic as the United Kingdom, and above all the trend is shifting: according to the EY Venture Capital Barometer, across the entire French Tech sector, 7.4 billion euros were raised in 2025 through 618 deals, representing a 5% decline in value. Without Mistral, the decline in investment value in French Tech would reach 26%, highlighting the structural fragility of the French market.

The UK’s ability to attract capital is largely due to the depth of its financial market, the presence of large specialized funds, a particularly well-developed entrepreneurial culture, and highly advantageous R&D tax incentives. Recently revised, these remain more ambitious than our Research Tax Credit and specifically target small businesses: for example, with the “super-deduction,” a tech SME investing £100,000 in an R&D project can deduct £230,000 from its taxable profit. International investors also view the country as a prime gateway to global technology markets. Finally, British universities, notably Oxford, Cambridge, Imperial College London, and University College London, constitute an exceptional pool of research and talent in disciplines related to artificial intelligence.

According to an industry study, the UK’s AI ecosystem was estimated to be worth approximately $230 billion in 2025, more than the French and German ecosystems combined. This gives the UK a significant advantage in creating new companies, developing AI software, and commercializing innovative solutions.

In France, a strategy focused on infrastructure and technological sovereignty

Faced with Britain's lead in entrepreneurship and startup financing, France has pursued a different path, focusing on the infrastructure required to support the next generation of AI applications. This approach is based on the idea that control over computing power, energy, and data centers will be a key factor in sovereignty and competitiveness in the years to come.

This ambition was showcased at the Summit for Action on Artificial Intelligence held in Paris in February 2025. On that occasion, President Emmanuel Macron announced over 109 billion euros in private commitments to finance AI infrastructure in France. These investments primarily concern the construction of data centers, the deployment of high-performance computing capabilities, and the development of platforms for training large AI models.

Since this announcement, several major projects have further bolstered this momentum, some of which were confirmed at the latest Choose France summit. Notably, the Japanese group SoftBank announced investments of 75 billion euros in data centers dedicated to AI. Other initiatives, such as the artificial intelligence “gigafactory” project near Paris, led by the French private equity fund Ardian (formerly AXA Private Equity) and the British sustainable data center specialist Verne (wholly acquired by Ardian), represent several billion euros in additional investment.

France also benefits from a significant structural advantage: its electricity mix, which is largely based on nuclear power. Data centers designed for artificial intelligence consume enormous amounts of electricity. The availability of abundant, largely carbon-free, and competitive energy is therefore a major asset for attracting international investment in digital infrastructure. Added to this, as Thomas Veyrenc (RTE) notes in a LinkedIn post, electric grid planning (which has identified priority areas for electrification) and the “fast-track” procedure that allows large data centers requiring very high electrical power (between 500 and 1,400 MW, equivalent to the power demand of a city of one million inhabitants) to be connected to the grid very quickly.

This strategy aims to make France a European hub for high-performance computing and AI model hosting, mirroring the role played by certain U.S. states in the global digital economy.

Building both stages of the rocket

The comparison between France and the United Kingdom is about more than a straightforward competition between two countries. Each is developing a specific model of leadership in artificial intelligence. The United Kingdom currently dominates high-value-added segments related to innovation, entrepreneurship, and startup funding. Its ecosystem is particularly effective at transforming research into fast-growing companies and attracting international investors.

France, for its part, is seeking to establish a strategic position in the infrastructure that will support the AI economy in the long term. The amounts announced for data centers and computing capacity now far exceed those seen in most other European countries. This focus could enable France to become a key player in hosting and training future artificial intelligence models.

The French approach, centered on computing centers, was recently validated and elaborated upon by Arthur Mensch, the young CEO of Mistral AI, during his hearing before the National Assembly. He explains that AI is not an intangible activity: it relies on a very concrete production chain that mobilizes energy, infrastructure, chips, talent, and capital. Its production involves transforming large quantities of electricity and investments into “tokens” (fragments of text, code, numbers, or images), which have become the basic economic unit of the AI economy. AI is becoming the central infrastructure of the digital world, just as the cloud was in the past. According to Arthur Mensch, the production of tokens resembles an industrial activity, even the heavy industry of yesteryear. Like any heavy industry, it raises issues of sovereignty, because relying on tokens produced elsewhere amounts to depending on others for an increasing share of one’s productivity and competitiveness.

The French strategy makes sense, and nuclear power is a valuable asset. In the medium term, moreover, the two approaches could prove complementary. British companies specializing in software and AI models will need massive computing infrastructure, while data centers located in France will need a dynamic ecosystem of innovative companies to fully utilize their capabilities.

Yet the success of Mistral AI suggests that France can—and should—aim to combine both approaches. After all, the Paris region also boasts remarkable scientific density, with Paris-Saclay University and Hi! PARIS on one side, and PSL University, Sorbonne University with SCAI, and Inria on the other. We should also mention initiatives such as the Kyutai laboratory, founded in 2023 by Xavier Niel (Iliad), Rodolphe Saadé (CMA CGM), and Eric Schmidt (former CEO of Google), which develops open-source models.

AI thus presents an opportunity to accelerate the development of a deep tech ecosystem, comprising startups that leverage disruptive innovations emerging from university laboratories. This development, launched a quarter-century ago by the Allègre Law prior to the decisive step of the Future Investment Program in 2009, reflects the government’s gradual adoption of the innovation-driven growth model described in the work of Philippe Aghion.

Yet several structural weaknesses continue to limit France's momentum: a venture capital market that has not yet reached scale (one reason being the incomplete development of the European capital market), few exits and thus little reinvestment, and an entrepreneurial culture that remains underdeveloped among researchers and within research and higher education institutions. Despite the spectacular progress of the French deep tech ecosystem since the 2010s, several challenges remain. The first is scaling up: many innovative companies struggle to become global leaders and are often acquired before reaching critical mass. In this scale-up segment, which requires significant investment with no promise of short-term returns, capital needs remain substantial, particularly in the industrial, quantum, space, nuclear, and biotechnology sectors. Development cycles are long and risky, delaying market access and requiring patient investors. Transforming scientific discoveries into profitable products remains a challenge, as demonstrated by biotech firms Genset and DBV Technologies. Some startups still face difficulties with industrialization and commercial adoption, such as Navya (autonomous vehicles) or, more recently, Ynsect (animal proteins). The most capital-intensive sectors remain exposed to technological uncertainties and sometimes immature markets. In the nuclear and space sectors, dependence on government decisions, regulations, and institutional support remains strong. French companies also suffer from a historical lack of international ambition and a reduced ability to pivot quickly in the face of global competition. Finally, the failures of projects like Astrid serve as a reminder that even strong scientific backing does not guarantee industrial success. The central question therefore remains France’s ability to sustainably transform its scientific excellence into global industrial champions.

The French AI ecosystem must be understood within this broader context, which helps explain why Paris still trails London in several key indicators. Yet AI also offers a unique opportunity to accelerate progress, not least because unprecedented amounts of capital are currently flowing into the sector.

While, as Arthur Mensch points out, the window of opportunity will not remain open for very long, the game is wide open today: in an interview with La Tribune Dimanche, he notes that when it comes to AI, talent and expertise are “abundant in Europe,” and adds: “What is true, however, is that the European technological ecosystem is newer, since the Americans got started before us. They then deployed their technology in Europe, a bloc that for a long time saw itself primarily as a continent of consumers. Hence the disconnect: trade flows originating in Europe are reinvested in research and development in the United States or China, rather than here at home.”

Catching up remains entirely possible; the challenge now is to accelerate. France has taken a significant lead in building the industrial and energy foundations of European artificial intelligence. It now needs to build the second stage of the rocket.

This first requires fostering the emergence of more startups from research labs by further streamlining the pathways between public research and business creation. This then requires more abundant capital to support companies during their scaling-up phase, to prevent them from being acquired prematurely or forced to go abroad for financing. Finally, it requires more dynamic demand, driven by both large corporations and public procurement.

The stakes extend far beyond the technology sector alone. Artificial intelligence is gradually becoming a general-purpose technology, comparable to electricity or the Internet, that will transform the entire economy. It is likely that the countries that master the infrastructure, models, and applications will capture a disproportionate share of the productivity gains associated with this revolution.