When EU leaders met at a medieval fort this week to debate how you can ramp up Europe’s financial energy, they had been confronting a stark actuality: Europe is lagging behind the innovation race with China and the US.
In its 2026 competitiveness report, the European Fee admitted that Europe “dangers shedding its edge within the race for innovation”. It pointed to labour shortages, issues to scale up, low numbers of patent purposes, and restricted R&D spending (under the three p.c of GDP goal).
Europe is robust in AI analysis papers, however the US dominates in AI corporations (with prime corporations like OpenAI, Google, or Nvidia) due to its Large Tech sector and huge enterprise capital flows, whereas China dominates in fast deployment and scale, aiming to grow to be an AI superpower by 2030.
A tortoise vs two hares
Total, it’s estimated that the EU has 5 p.c of the world’s complete AI computing energy, in comparison with the present US, which has 74 p.c of world high-end AI compute, and China, which has 14 p.c.
To catch up, the EU has been quickly making an attempt to broaden its AI firepower, with the fee pushing cash and power in direction of growing an EU-wide AI uptake technique and a number of AI factories throughout the bloc — whereas reopening the digital rulebooks by means of the digital omnibus to make information extra accessible.
Different proposals embody the Innovation Act and Scaleup Europe Fund to assist start-ups and fast-growing companies.
The EU set a goal of mobilising €20bn in AI funding between 2018 and 2020, and €20bn yearly within the decade that adopted.
The fee pledged to boost its personal contribution to €1.5bn for 2018-2020 and €1bn per yr for 2021-2027.
However watchdogs have warned that progress has been sluggish and uneven.
France and Germany have unveiled the biggest nationwide funding plans, whereas Italy, the Nordics, Spain, and Greece are additionally ramping up extra modest funding deal with supercomputing and SMEs.
In opposition to this backdrop, why has the bloc been struggling to date?
Late begin, sluggish scale-up, fragmented strategy
Europe finds itself on this place as a result of it began scaling its AI capabilities a lot later than different areas, such because the US, which surged within the post-2010s, and China’s state-backed plans from 2017.
After which, when its ramp-up started, the 27-nation bloc didn’t have enough funding within the know-how, which has introduced scaling and distribution struggles to the continent.
However the late begin doesn’t imply Europe has been incapable of growing its personal AI.
“There’s a robust analysis background and engineering expertise pool in Europe. We do have some corporations even growing these foundational [AI] ]fashions,” defined Leevi Saari, doctoral researcher on the College of Amsterdam learning the political financial system and synthetic intelligence.
Mistral AI, a French AI firm with a valuation of €11.7bn, is commonly pointed to as a European champion.
The problems come from scaling European developments.
“What Europe doesn’t have, which is the important thing downside in all of this, is that we’re missing the commercial platform or kind of coordinated scalable platform that exists within the US or China,” Saari added.
Within the US, for example, AI growth has been superior by Large Tech, which has entry to huge quantities of coaching information that corporations like Google have been hoarding for years, and may distribute their tech throughout monumental platforms and to their companions.
Europe at present doesn’t have the hyperscalers, information repositories, and centered industrial flexibility to maneuver and develop as shortly as US or Chinese language corporations.
With a part of this difficulty additionally being the piecemeal strategy the bloc has taken in direction of the know-how.
The fee has been making an attempt AI-related initiatives since 2018, however its strategy to coordinating with member states on growth and funding has been criticised as fragmented and ineffective.
The European Court docket of Auditors launched a report in 2024 that claims: “Greater than 5 years after the primary plan, the framework for coordinating and regulating EU funding in AI remains to be a piece in progress.”
The report factors to the shortage of governance instruments and to the absence of coordinated, particular funding and growth targets.
Bengüsu Özcan, AI governance researcher at Brussels-based think-tank Centre for Future Generations, defined to EUobserver: “capital, computing energy, and expertise are nonetheless fragmented throughout member states and areas.”
“Particular person member states can not attain the brink of assets wanted to construct the large-scale compute clusters, entice world-class expertise, and maintain the long-term investments that cutting-edge AI methods demand,” she added.
Market dynamics
One other a part of the AI downside for the bloc is the shortage of funding to genuinely compete with the massive world gamers.
“We [the EU] wish to get away by not investing sufficient and nonetheless being aggressive. However with our present funding degree, we must always not likely be the US and China, as a result of they function at one other degree, if you happen to ask me,” stated Robert Praas, information and AI scientist at Brussels-based thinktank CEPS, advised EUobserver.
The European Fee launched InvestAI in 2025, a public-private funding initiative aiming to have €200bn in direction of AI and AI infrastructure.
Examine this to OpenAI, which alone needs to speculate $500bn [€422bn] over the subsequent 4 years into AI infrastructure.
“I don’t suppose it [EU commission] has the assets to essentially change the panorama as a result of it has a restricted funds in comparison with American enterprise capital mainly,” stated Praas.
What Europe is missing is robust, non-public, enterprise capital (VC) funding.
A lot of Silicon Valley AI growth occurs by means of VC, the place non-public cash invests, typically in return for firm fairness. The US invested €1.33 trillion in VC between 2020 and 2025, 34 p.c of which went to AI.
The EU invested solely €252bn in VC over the identical interval.
To assist shut the non-public capital hole, the fee is wanting into adapting the bloc’s enterprise capital capabilities to push extra non-public funding into the know-how.
Nevertheless, the lifecycle of a venture-funded firm is normally to construct till you possibly can promote to Large Tech or go public, with the intention of constructing large returns for traders.
This might increase issues for European enterprise capital-funded corporations, as the inducement to maneuver to the US market is at present stronger than staying in Europe.
“In case you don’t take into consideration the entire life cycle of these sorts of investments, it’s simple to make, once more, counter-productive errors the place we then create corporations that may ultimately be purchased, or offered to the US,” added Saari.
However what’s the top aim?
The extreme leveraging by the EU to grow to be aggressive in AI comes because the bloc is rattled by worrying competitiveness experiences and by rising hostility from longtime know-how companions such because the US.
However because the bloc scrambles to develop its personal technological sovereignty, the researchers famous that the continent’s AI targets appear huge.
“We would like Europe to be one of many main AI continents, and this implies embracing a lifestyle the place AI is in every single place,” stated fee president Ursula von der Leyen on the Paris AI summit final February.
The InvestAI initiative, for instance, goals to assist create infrastructure for next-generation fashions that purpose to spice up the whole lot from medical developments, scientific analysis, and financial competitiveness.
And the specialists are sceptical that such a strategically big strategy, and attempting to immediately compete with giants like OpenAI on normal fashions, goes to create actual developments in the way in which the Brussels govt hopes. As a substitute, they recommend extra strategic readability.
Saari notes that there’s a have to incentivise the selection of European alternate options over overseas fashions.
At the moment, OpenAI is actively funding initiatives to coach 20,000 European small companies and to signal member state governments as much as its platform — demonstrating the extreme overseas effort to steal European clients.
“If Europe needs AI sovereignty, no matter meaning, extra essential is to persuade the nice folks in German trade [for example] to decide on Mistral over OpenAI,” stated Saari.
And Praas believes tailoring funding and growth to explicit sectors, relatively than the whole lot without delay, is a stronger strategy.
“If we are saying, hey, healthcare is so essential and we expect this know-how will help treatment extra folks, assist extra folks, let’s choose this one.
“Let’s create a programme the place folks actually incentivise going into this one,” stated Praas.
Suggesting, closely specializing in a few sectors at a time and incubating growth may very well be a means that the EU can compete extra successfully.
“If that’s the case, yeah, then the long run will be very brilliant,” he added.
