The European Digital Sovereignty Race: The Competitive Landscape of AI and Quantum Computing from EPO Patent Data

On the global technology map of 2025, Europe stands at a delicate juncture. According to the latest special insight from the European Patent Office (EPO) technology dashboard, computer technology remains the strongest engine for European patent applications, but the growth momentum in AI, the explosive growth in quantum computing, and the relative slowdown in semiconductors together paint a competitive picture far more complex than simply "leading" or "lagging behind."

Patent data becomes a barometer of digital sovereignty

Patent data is not only a record of technological innovation, but also strategic intelligence on national and regional competitiveness. According to EPO data, applications in the field of computer technology grew by 6.1% year-on-year in 2025, firmly ranking first. Among them, AI applications grew by 9.5%, while quantum computing, despite a smaller base, grew at a rate as high as 37.9%. Behind these figures lies intense strategic positioning by various innovators around next-generation computing infrastructure.

The AI race: Europe's "lead" and America's "acceleration"

In the AI field, Europe still holds a slight lead globally, accounting for more than one-third of all EPO AI applications. However, this leading position has begun to show signs of fatigue—European applications grew by only 2.6% in 2025, while the United States grew by 26.4%. Europe itself is also experiencing divergence: German applications fell slightly by 1.1%, but France grew by 19.0% and the United Kingdom by 28.8%. Germany remains the absolute mainstay of European AI innovation, but the rapid catch-up by France and the UK indicates that Europe's innovation gradient is undergoing structural change.

Meanwhile, Asia has not been absent. South Korea grew by 16.5%, China by 3.9%, and Japan fell by 8.6%. The "three-way race" among the three major innovation poles of Asia, Europe, and the United States is shifting from overall scale to localized battles over specific technology routes.

Subfield differentiation: Europe's strengths and weaknesses

AI is not a homogeneous track. EPO data shows that neural network-related applications based on biological models account for more than half of all AI applications. Europe's share in this field has remained stable at around 35% since 2017, but declined slightly in 2025 as the United States regained momentum. Image and video recognition is the second largest area of AI applications, growing rapidly since 2021, with Europe significantly ahead of the United States, while the United States is accelerating its approach. In the fields of machine learning and pattern recognition, Europe has long lagged behind the United States. The share of machine learning peaked in 2023 and then declined, stabilizing at around 30%; pattern recognition accounts for less than 4% of AI applications, and the United States has once again overtaken Europe.

This differentiation across subfields suggests that Europe needs to increase investment in fundamental algorithms and model architectures, while in application layers (such as medical imaging and autonomous driving visual recognition) it may build stronger barriers.

Quantum computing: accelerating deployment, but competition is intenseQuantum computing is the fastest-growing field in the 2025 EPO data. Despite a low base, global filings grew by 37.9%. European innovators grew by 22.1%, the US by 43.3%, and Japan achieved a 170.8% leap from a low base. Europe still leads in the total number of quantum computing patents, but the rapid growth of the US means that Europe must move faster to convert research advantages into patentable technical assets and build an ecosystem that integrates quantum computing with AI and semiconductors.

Semiconductors: Europe “Still Strong” but “Stalling”?

Within the broader framework of computer technology, semiconductors are the physical foundation of AI infrastructure. EPO insights point out that European applicants remain strong in semiconductors but are “losing momentum.” This may be a snapshot of the current state of Europe’s semiconductor industry — despite giants such as Infineon and STMicroelectronics, Europe faces denser patent portfolios from the US, South Korea, and Taiwan in advanced process nodes and AI-specific chips. This signal may also reinforce the urgency of the EU’s Chips Act: beyond mature process nodes, Europe needs to build new advantages in chip design, packaging, and energy-efficiency technologies for the AI era.

The Deep Coupling of AI, Energy, and the Green Transition

AI is not just a race in computing power; it is also an energy race. Citing data from the International Energy Agency (IEA), the EPO notes that large AI data centers can consume as much electricity as 100,000 households, while the largest data center currently under construction could consume power equivalent to 2 million households. Demand on this scale is driving innovation in two directions: first, reducing energy consumption per computation through more efficient chips, quantum computing, and other means; and second, using AI to optimize renewable energy forecasting, grid dispatch, and energy-efficiency management — thereby “using AI to solve AI’s energy problem.”

This coupling of energy and AI is precisely the key intersection where the European Green Deal meets industrial competitiveness. If Europe can leverage its existing strengths in image recognition, industrial automation, and other fields, and layer on energy-efficiency innovation, it will have the opportunity to carve out a unique position on the “green AI” track.

Two Samples from the Innovation Ecosystem

Behind the EPO data are specific innovators. Norwegian inventor Esben Beck developed an underwater robot based on AI image recognition and lasers to remove sea lice from salmon farms; a Lithuanian team built an AI platform to design industrial enzymes from first principles, serving biomanufacturing, precision medicine, and other fields. These two cases show that Europe’s innovation vitality comes not only from large enterprises but also from small teams and independent inventors. How to turn these “fringe innovations” into scaled advantages through the patent system, venture capital, and industrial collaboration is a challenge that European policymakers must address.

A Key Variable for Strategic AutonomyFrom the EPO's insights, it is clear that Europe's overall competitiveness in AI, quantum computing, and semiconductors has not collapsed, but it is entering a critical stage of "advance or fall behind." The strong rebound of the United States, Asia's catch-up in certain areas, and imbalances within Europe call for Europe to strike a better balance between AI regulation (the AI Act) and technology investment, and to draw clearer boundaries between open cooperation and strategic autonomy.

Patent data does not offer direct answers, but it provides Europe — and all observers focused on European competitiveness — with an evidence-based map. In the years ahead, how Europe uses this map to make choices will determine the real substance of its digital sovereignty.