Some races are won or lost in the first moments after the starting whistle, so let’s get this out of the way: as a whole, Europe is not competitive with the U.S. or China in developing the high-scale, foundational large language models (LLMs) on which the AI economy depends.
The continent’s sole noteworthy LLM, France’s Mistral, is the exception that proves the rule, and still substantially smaller than those of global market leaders like OpenAI, Google, Meta, Deepseek or Anthropic. The sums being invested into these American and Chinese models make catching up unlikely.
Does this mean that Europe has lost its chance to benefit from the AI revolution on equal terms with the U.S.?
Not necessarily. The value of AI mostly manifests in how firms use the technology, says Matthias Tauber, who leads Boston Consulting Group’s operations in Europe, Middle East, South America, and Africa. “When it comes to AI adoption, we don’t see a difference between European or U.S. companies. Whether they will be winners, yes or no, will be determined by who drives adoption faster,” he tells Fortune.
Dominic King, EMEA research lead at Dublin-based consultancy and IT firm Accenture, agrees: “European companies are well-positioned to add value by building applications on top of general-purpose U.S. models.”
In other words, it’s still all to play for.
Which European companies are ahead?
When it comes to AI adoption, Europe still has its work cut out for it. According to the European Parliament, only 13.5% of EU companies were using AI as of last year. While that’s no doubt increased substantially since, it’s a far cry from Europe’s 75% target. It’s also likely to be well behind the U.S., with McKinsey estimating a 45-70% transatlantic adoption gap in the same year.
Zoom in, however, and you’ll see a more nuanced picture, with many European firms at least keeping up with their global competitors.
“When it comes to AI adoption, we don’t see a difference between European or U.S. companies. Whether they will be winners, yes or no, will be determined by who drives adoption faster.”
Matthias Tauber, Head of BCG Europe, Middle East, South America and Africa
Much depends on the size of the business. Accenture research on Europe found a clear relationship between the strength of an organisation’s AI capabilities, such as its talent and data governance, and its AI deployment. Larger companies are “typically able to invest more, have stronger change management skills and benefit from larger datasets,” King says.
Which companies are ahead also depends on their sector. Alongside the obvious candidates like IT, many of Europe’s leading industries—like automotive, biopharma, fintech and aerospace—are among those where AI significantly impacts core activities, rather than just supporting functions. This makes them both ripe to benefit from AI deployment, and vulnerable to external disruption of the kind already playing out in electric vehicles.
That blend of threat and opportunity has made firms in these sectors more likely to actively lean into the new technology. “Here we see early adopters boosting productivity with AI, for example, by accelerating drug discovery, conducting more accurate simulations and improving product design,” adds King.
Accenture itself, while best understood as a multinational with European headquarters rather than as a distinctly European company, is among those early adopters. In 2023, Accenture announced it would set aside $3 billion to integrate AI internally and to become experts on it for its clients, per previous Fortune reporting.
The firm booked $4.1 billion for GenAI work, and $1.8 billion in revenue, as of its Q3 earnings call in June, with embedded AI, deep industrial knowledge and energy efficiency emerging as key themes. It is aiming to build an 80,000-strong data and AI workforce by 2026, having already hit 75,000.
Larger companies are “typically able to invest more, have stronger change management skills and benefit from larger datasets.”
Dominic King, EMEA research lead, Accenture
Schneider Electric is another European company going big on AI. The industrial technology and energy management group generated over €100 million (around $116.9 million) in business value from embedding AI into its operations, Gwenaelle Avice Huet, its executive vice president of operations in Europe, tells Fortune. That figure, which actually dates back to as early as 2022, is a result of cost savings and operational efficiencies it made via its “self-healing” supply chain platform.
The French multinational uses AI in its supply chain, financial advisory and customer service. “Our internal Jo-ChatGPT platform enables employees to securely leverage generative AI, boosting productivity and creativity while maintaining data integrity,” Avice Huet adds. Externally, AI is also used in Schneider Electric’s flagship products, such as energy management and industrial automation.
The main way that Schneider Electric benefits from the AI boom is more direct, though, due to its role as a leading global supplier of electrical components used in data centers, alongside others like the Netherlands’ ASML, a key technology supplier for advanced semiconductor manufacturers.
To give a sense of the size of the market they are supplying, in the EU alone €100 billion in data center investments are projected by 2030, according to the European Data Centre Association, although this is likely to be substantially lower than the equivalent in the U.S., which McKinsey estimates will alone receive around 40% of global data center investment this decade.
Some of this investment is coming from companies that you wouldn’t normally call tech firms, with EU businesses such as Lidl’s parent company, Schwarz Gruppe, eyeing their own data centers, partly from a desire to reduce Europe’s dependence on American capabilities.
Not everyone is proving so enthusiastic, however. As in other countries, there are also prominent sectors of the European economy that tend to lag in AI adoption, such as utilities and telecommunications—ironically, sectors that themselves underpin the rollout of AI. King explains that these struggle with fragmentation, access to capital, and weak AI capabilities due to low AI literacy and a lack of concrete use cases with clear return on investment.
A double infrastructure gap
Despite some stragglers, the big picture is of soaring demand for AI, but even with the vast sums being invested in European data centers, supply is still struggling to keep up. As a result, infrastructure risks becoming a critical bottleneck, making AI more expensive and slower to use. Data center vacancy rates—a measure of their available additional capacity—are at an all-time low on the continent.
AI adoption is also likely to come up against another infrastructure bottleneck, in the energy system. Data centers use substantial electricity—Goldman Sachs predicts they could add 40-50% to Europe’s power demand over ten years.
This causes two problems. First, the additional burden on the grid will apply upward pressure on Europe’s high energy prices, which already weigh on industrial competitiveness. Second, if Europe’s energy infrastructure investments can’t keep up with data center demand, then it risks constraining AI adoption for European businesses.
It’s not just the lack of power per se. Data centers depend on an uninterrupted energy supply, but the product they facilitate creates demand spikes that make outages more likely. If there’s too much volatility, it can impede their operations, add costs and disincentivize further investment.
“If you’re a data center operator, you’re sat in the middle of double uncertainty, with more volatility coming in on the demand side and more volatility on the energy market side,” says Jade Batstone, cofounder and CEO of Zendo, a startup helping data centers become more energy efficient.
The danger for Europe’s competitiveness is that its economy could fall relatively further behind on both AI and energy prices, in the absence of accelerated, simultaneous investment into both sets of infrastructure.
It would be a mistake to see AI only as a problem for the energy sector, however. It can also be part of the solution. The International Energy Agency (IEA) projects that AI could unlock an additional 175 gigawatts of global energy capacity simply by improving the efficiency of grids, which is more than just a marginal efficiency gain: it’s higher than the total projected global energy demand for data centers by 2030, and five times more than Europe’s 2030 projected power demand.
Standing firm on going green
This points to the one area where Europe has something of an advantage over the U.S.—the intersection between data centers and renewable power.
Europe has “a strong legacy” in data centers, clean technology and manufacturing, meaning its competitive edge lies in “building the resilient, sustainable infrastructure that powers AI,” argues Avice Huet, pointing to Schneider Electric’s partnership with Nvidia on AI-native data center designs.
“While high energy costs may weigh on Europe’s competitiveness today, particularly in energy-intensive industries, smart deployment of AI combined with the continent’s leadership in renewables technologies such as offshore wind could help reduce both emissions and costs in the long-term,” King adds.
Such an outcome is particularly appealing for businesses that are committed to both AI and decarbonization, with large tech firms such as Google setting the bar with commitments to be fully powered by renewable energy by 2030. Indeed, BNP Paribas notes that most hyperscalers favor renewables even on economic grounds alone, owing to the lower operational costs from solar, geothermal and wind.
But fully renewable data centers may not be so straightforward to achieve, notes Zendo cofounder and COO Drew Barrett: “You’re going to really struggle to do that in grids that haven’t deeply decarbonized already.”
This is where Europe’s advantage comes in. While no one has hit the full decarbonization bar yet, renewables did generate 50% of all electricity used in the European Union last year, per the IEA—comfortably the highest of the major economies. Brussels has also set the goal for data centers to be climate neutral by 2030, requiring them to report on energy consumption, how much of that is renewable, and water usage.
While some may see such regulation as an additional barrier to investment, Avice Huet argues that “decarbonization is not a constraint on competitiveness; it is central to Europe’s ambitions for growth and industrial strength.”
This position mirrors the EU’s Clean Industrial Deal, a strategy to build a competitive niche in clean technologies, which could extend to ‘green’ data centers—especially as the U.S. looks to fossil fuels to power its computing needs, following President Trump’s AI action plan, which was notably silent on renewables.
But while European lawmakers’ studied focus on consumers over businesses has resulted in world-leading laws on data protection and sustainability, Tauber says it has still complicated the private sector’s ability to actually compete. Given the complexity and fragmentation of EU legislation, which is interpreted differently across member states and sits on top of multiple layers of domestic law, Europe should deregulate, he says.
There has been some progress in simplifying regulations. The EU’s AI Continent Action Plan proposes streamlined permitting for data centers that meet energy and water efficiency standards, meaning green data centers are preferentially incentivized. Avice Huet sounds an optimistic note: “With a continued focus on cutting the red tape, electrification, digitalization, and grid modernisation, Europe can emerge stronger, more resilient, and more competitive on the world stage.”
