When the European Commission unveiled its Tech Sovereignty Package in June, the message was clear: Europe wants to reduce its dependence on non-European providers of chips, cloud infrastructure and artificial intelligence – or, more precisely, on a handful of dominant American technology companies.

Looking at the numbers, that concern is reasonable.

Around 70% of Europe's cloud market is controlled by Amazon Web Services, Microsoft Azure and Google Cloud. European providers have seen their combined market share fall from 29% in 2017 to just 13% five years later . Europe accounts for roughly 5% of global AI computing capacity compared with around 75% for the United States , while European organisations spend an estimated €264 billion each year on U.S. digital services.

These numbers are usually discussed as a technology gap. But there is a bigger question underneath them. As cloud computing and artificial intelligence become embedded in critical industries, including energy, the issue is no longer simply who develops the technology, but how much strategic leverage countries retain over the digital infrastructure their economies increasingly rely on. For Europe, the priority is to ensure that dependence does not become vulnerability.

America Didn't Just Build Better Technology. It Built The Ecosystem.

The United States' digital leadership did not emerge overnight.

While much of Europe’s post-war competitiveness remained anchored in advanced manufacturing, industrial exports and long investment cycles, the United States increasingly concentrated on software, intellectual property and platform businesses. Deep capital markets, world-leading universities and a venture-capital ecosystem designed to tolerate failure created powerful feedback loops between talent, finance and innovation.

When the internet economy scaled, that ecosystem was already in place. 1

Without digital companies of comparable scale, Europe’s influence came increasingly through regulation. GDPR, the Digital Markets Act and the AI Act shaped how global technology companies operate. They did far less to create European alternatives. 2 3

Artificial intelligence has widened that gap. Since 2023, Amazon, Google, Meta and Microsoft have spent $1.1 trillion on AI infrastructure. This year, another $745 billion is expected to follow – capital few governments, let alone competitors, can match.

That is not to say Europe is absent from the technology stack. It has critical strengths: world-leading semiconductor capabilities through companies such as ASML, a world-class scientific and positioning in quantum computing , deep industrial know-how, strong research institutions and a growing generation of AI companies. The challenge is not a lack of innovation, but the ability to scale it through the wider ecosystem needed to support that growth. 4

Europe’s most prominent AI companies illustrate the point. Mistral, the French AI company often positioned as Europe’s answer to OpenAI, develops its models in partnership with Nvidia ; DeepL, the German translation company that has operated its own renewable-powered servers since 2017, entered a partnership with Amazon Web Services this spring; and Swedish AI startup Lovable, valued at $6.6 billion, relies on Anthropic and Google models and recently expanded its Google Cloud footprint fivefold. 5 6

These partnerships are not a weakness in themselves. Rather, they show how difficult it is for even successful European technology companies to operate independently of the infrastructure, computing power and foundation models controlled by a small number of global providers.

When Digital Dependence Reaches The Grid

That dependence matters far beyond the technology sector once the same capabilities become embedded in critical infrastructure. Energy offers one of the clearest examples. Modern electricity systems no longer rely only on power plants and transmission lines. They depend on software capable of forecasting renewable output, balancing electricity demand, predicting equipment failures and coordinating millions of distributed assets in real time.

Those capabilities already deliver measurable value. In the United States, the Midcontinent Independent System Operator is working with Microsoft to develop AI-enabled forecasting tools capable of helping manage more than 350 gigawatts of proposed generation waiting to connect to the grid. In Great Britain, Google's DeepMind models have improved solar forecasting for the National Grid Electricity System Operator by around 40% .

Data centers currently account for roughly 2% of EU electricity demand, but that share could rise to 7% or 8% by 2030. 7 As AI scales, computing is becoming an energy business as much as a digital one.

For Europe, that presents both a constraint and an opportunity. In several markets, data-center development is already outpacing grid expansion, with connection timelines stretching from seven to thirteen years. 8 At the same time, Europe has strong capabilities in grid technologies and in building and operating increasingly complex power systems. It has another structural advantage too: cleaner electricity. In 2024, the average carbon intensity of EU grid power was around 213 grams of CO₂ per kilowatt-hour, compared with 384 grams in the United States, and that gap is expected to widen as renewable deployment accelerates. 9

Cleaner electricity and grid resilience alone will not close Europe's gap in cloud infrastructure or AI investment. But as access to reliable, affordable low-carbon power becomes a bigger factor in where computing capacity is built, it gives Europe an asset that few regions can match.

Europe may never rival the United States on hyperscale cloud infrastructure. And it does not necessarily have to. Instead, it has an opportunity to become an attractive location for what is now being described as clean compute, provided it can continue expanding grid capacity, storage and flexibility quickly enough to support that ambition.

In other words, AI will shape how energy systems operate, while energy will help determine where AI can scale.

Sovereignty Is About Leverage, Not Self-Sufficiency

This two-way dependence helps explain why the Tech Sovereignty Package marks an important shift. For much of the past decade, Europe sought to regulate the digital economy. Now it is attempting, selectively, to build parts of it. Not all of it.

Replacing foreign providers across semiconductors, cloud computing, artificial intelligence and software has been estimated to cost roughly €3.6 trillion over the next decade. That’s nearly twice the EU’s proposed budget from 2028 to 2034, covering everything from agriculture to defense. 10 11

That price tag alone explains why Brussels is not pursuing full technological independence. Instead, the approach is more pragmatic.

The revised Chips Act strengthens semiconductor manufacturing. The Cloud and AI Development Act seeks to triple Europe’s data-centre capacity within five to seven years, while reserving only the most sensitive public-sector workloads – around 1% – for providers operating under European control. Open-source alternatives and maintenance receive €2 billion in targeted support, while the accompanying energy roadmap aims to integrate data centers into grid planning, encourage flexible operation and assess their energy use, water consumption, renewable supply and waste-heat potential. 12

The objective is therefore not technological self-sufficiency, but to strengthen Europe’s leverage within the global technology ecosystem. That means retaining credible alternatives in strategic areas, protecting its most sensitive capabilities and using strengths such as clean energy and industrial infrastructure to ensure that digital reliance does not become strategic vulnerability.

That approach also creates a different investment thesis.

The opportunity is unlikely to come from replacing today's technology leaders. It is more likely to emerge around the technologies and services needed to make the digital system more resilient.

Expanding Europe’s computing capacity will require far more than additional data centers. It will demand investment across power systems, grids, storage, cooling, cybersecurity and industrial software. 13 14 15

Rather than replicating the existing technology stack, value is likely to come from strengthening the infrastructure and capabilities around it.

The Next Layer Of The Energy Transition

The energy transition has always been about replacing one infrastructure with another. But the infrastructure now being built is as much digital as physical.

Wind farms, batteries and transmission lines generate, store and transport electricity. Software, cloud computing and artificial intelligence help determine how efficiently and reliably those assets operate together.

Europe does not need to recreate Silicon Valley or own every layer of the technology stack. Today’s challenge is different: ensuring that the digital layer underpinning its energy transition becomes a source of resilience rather than strategic vulnerability.

Because in the digital age of industrial competitiveness, the question will no longer simply be who produces the cleanest electricity. It will also be who has the capabilities to manage and optimize the system that delivers it. And that matters beyond energy. As digital systems become embedded across the economy, Europe does not need to build everything itself. But it does need enough capability and choice to remain competitive and resilient as those dependencies deepen.

1 https://commission.europa.eu/topics/competitiveness/draghi-report_en

2 https://netchoice.org/through-the-regulatory-fog-how-the-eus-digital-markets-act-is-harming-consumers-and-innovators/

3 https://publications.jrc.ec.europa.eu/repository/handle/JRC146878

4 https://www.thetimes.com/business/companies-markets/article/how-asml-became-europes-most-valuable-company-nwp7cl5hq

5 https://www.techpolicy.press/how-ai-keeps-europe-hooked-on-us-cloud/

6 https://euaipolicymonitor.substack.com/p/the-limits-of-building-european-ai

7 https://www.europarl.europa.eu/RegData/etudes/BRIE/2025/775859/EPRS_BRI(2025)775859_EN.pdf

8 https://www.europarl.europa.eu/RegData/etudes/BRIE/2025/775859/EPRS_BRI(2025)775859_EN.pdf

9 https://ember-energy.org/latest-insights/global-electricity-review-2025/major-countries-and-regions/

10 https://cepa.org/article/digital-sovereignty-can-europe-afford-it/

11 https://commission.europa.eu/strategy-and-policy/eu-budget/long-term-eu-budget/eu-budget-2028-2034_en

12 https://digital-strategy.ec.europa.eu/en/policies/eu-tech-sovereignty

13 https://www.iea.org/commentaries/overcoming-energy-constraints-is-key-to-delivering-on-europe-s-data-centre-goals

14 https://iea.blob.core.windows.net/assets/40a4db21-2225-42f0-8a07-addcc2ea86b3/EnergyandAI.pdf

15 https://www.oecd.org/en/publications/oecd-digital-economy-outlook-2024-volume-1_a1689dc5-en.html