AI in Europe and the world

1  2  3

1 Is there a European AI?

1.1 Key European AI Companies and Models

1.2 The European Approach: Regulation and Trust

1 3 Investment and Infrastructure

1.4 Public and Research Initiatives

2 Which European countries are developing their own AI?

2.1 Key European Countries Developing Domestic AI:

2.2 Key Trends and Initiatives:

2.3 Top European AI Hubs by Startups:

3 Which countries are developing their own AI? 

3.1 Key Nations Developing Sovereign AI:

3.2 Regional and Smaller Initiatives:

 

1 Is there a European AI?

Yes, there is a rapidly growing European AI ecosystem

 that is actively developing, regulating, and deploying artificial intelligence, aiming for "technological sovereignty" to reduce reliance on US and Chinese tech giants. 

While major US firms currently dominate, Europe has established significant homegrown AI alternatives, research capabilities, and a unique regulatory framework. 

Here is a breakdown of the European AI landscape:

1.1 Key European AI Companies and Models

Several European companies are building "sovereign" AI, often focusing on open-weight models, data privacy, and B2B (business-to-business) applications: 

Mistral AI (France): A leading startup founded in 2023 that creates highly efficient, open-weight large language models (LLMs).

Black Forest Labs (Germany): Known for the FLUX family of generative image models, which offer high-fidelity, open-source alternatives to models like Midjourney.

DeepL (Germany): A prominent AI-powered translation and language tool widely used for its accuracy and data security.

Aleph Alpha (Germany): Focuses on sovereign, secure AI for enterprise and government applications.

ElevenLabs (UK): A major player in AI voice generation and audio, reaching unicorn status in 2024.

BLOOM (BigScience): A massive, open-source multilingual language model developed by a coalition of over 1,000 researchers, coordinated by Hugging Face (founded in France). 

1.2 The European Approach: Regulation and Trust

The European Union (EU) has established the EU AI Act, the world's first comprehensive, legally binding, risk-based regulation for artificial intelligence. 

Goal: To foster "trustworthy AI" that respects EU values and fundamental rights.

AI Office: Established in 2024 to oversee the implementation of the Act, particularly for general-purpose AI (GPAI).

Safety & Ethics: The approach focuses on limiting "unacceptable risk" (e.g., social scoring) and ensuring transparency. 

1.3 Investment and Infrastructure

To bridge the gap with the US, Europe is investing heavily in AI infrastructure: 

AI Factories & Supercomputing: The EU is launching "AI Factories" to provide supercomputing power (like LEONARDO and LUMI) to startups and researchers for training large models.

GenAI4EU: A program to boost the adoption of generative AI across 14 key industrial ecosystems.

"InvestAI": Initiatives aimed at mobilising billions in private investment to support European AI companies. 

1 4 Public and Research Initiatives

RAISE (Resource for AI Science in Europe): A virtual institute designed to support AI research for scientific breakthroughs.

National Models: Several countries are developing their own sovereign, language-specific models, such as PLLuM (Poland), Apertus (Switzerland), GPT-NL (Netherlands), and Amalia (Portugal). 

In summary, Europe is not just a consumer of AI but a significant developer of it, with a focus on high-quality, regula

<<

2 Which European countries are developing their own AI?

Several European countries are actively developing their own sovereign artificial intelligence capabilities to reduce dependence on US and Chinese technologies, driven by initiatives like the EU's "AI Factories" and national investments in large language models (LLMs).  

2.1 Key European Countries Developing Domestic AI:

France: Emerging as a leader in foundational models, largely due to Mistral AI, which has developed its own high-performance LLMs. Other notable developments include Poolside (GenAI software) and H-Company.

Germany: Focused on industrial AI, B2B, and sovereign open-source models. Aleph Alpha is a major developer of explainable, secure LLMs for European industry. The government has also launched the SOOFI (Sovereign Open Source Foundation Models) initiative.

United Kingdom: A top European hub with companies like Wayve (autonomous driving) and Synthesia (text-to-video). The UK is investing in its own supercomputers to enhance domestic capabilities.

Spain: Developed Alia, a project focused on building LLMs in Spanish, Catalan, Basque, and Galician for domestic use.

Switzerland: Launched Apertus, a multilingual, open-source language model developed by the Swiss AI Initiative, specifically designed for customized, sovereign AI.

Poland: Developed the PLLuM (Polish Large Language Model) designed specifically to handle Polish language nuances.

Netherlands: Working on GPT-NL, a sovereign, open-source model designed for the Dutch language and culture.

Portugal: Developing Amalia, a sovereign, open-source LLM for public administration and scientific research.

Bulgaria: Through the Institute for Computer Science, Artificial Intelligence, and Technology (INSAIT), developed BgGPT. 

2.2 Key Trends and Initiatives:

EU AI Factories: As of October 2025, 19 "AI Factories" are being deployed across 16 member states, including Finland, Germany, Greece, Italy, Luxembourg, Spain, Sweden, Austria, Bulgaria, Poland, Slovenia, Czech Republic, Lithuania, Romania, Belgium, and others. These facilities provide supercomputing power and infrastructure to train local models.

Sovereign AI Focus: Many projects are intentionally open-source and trained on European data to ensure compliance with privacy laws and reduce reliance on foreign tech giants.

Key Sectors: Development is heavily focused on industrial applications (Germany), autonomous vehicles (UK), and national language models (Spain, Poland, Netherlands).  

2.3 TopEuropean AI Hubs by Startups:

United Kingdom (334+ startups)

Germany (167+ startups)

France (135+ startups)

Estonia (High concentration of AI startups per capita) 

<<

3 Which countries are developing their own AI? 

Several countries are developing their own sovereign AI—indigenous large language models (LLMs) and AI infrastructure—to reduce reliance on US and Chinese technologies and ensure data sovereignty. 

3 1 Key Nations Developing Sovereign AI:

China: Pursuing a "national self-reliance initiative" to develop indigenous LLMs (e.g., Baidu's ERNIE, Alibaba's Tongyi, Huawei's PanGu) following US sanctions.

United States: Continues to lead in AI, with massive investment and major, private-sector-led development.

United Arab Emirates (UAE): Developed the "Falcon" series of open-source LLMs at the Technology Innovation Institute (TII).

Saudi Arabia: Investing in the "ALLAM" Arabic-first LLM through the Saudi Data & AI Authority (SDAIA) and the Public Investment Fund.

India: Launched the $1.25 billion "IndiaAI Mission" to build native infrastructure, compute, and models like Krutrim and Sarvam AI.

Japan: Developing domestic LLMs, with RIKEN and Tokyo Institute of Technology training models on the Fugaku supercomputer, aiming to improve Japanese language processing.

South Korea: Backing five local "AI champions" (LG, SK Telecom, Naver, NCSoft, Upstage) with a ₩530 billion initiative to create native Korean-language AI.

France: Championing European "digital sovereignty" by funding Mistral AI and supporting open-source models like BLOOM, while upgrading the Jean Zay supercomputer.

Germany: Focused on Sovereign Open Source Foundation Models (SOOFI) and backing Aleph Alpha for industrial, secure AI.

United Kingdom: Utilizing a National Foundation Model Taskforce to build "BritGPT" and investing heavily in compute infrastructure.

Canada: Implementing a $2 billion+ "AI Sovereign Compute Strategy" to build local infrastructure, with a focus on supporting startups like Cohere.

Russia: Pursuing "sovereign AI" through Sberbank's GigaChat and Yandex's YaLM 2.0 to ensure independence from Western platforms.

Taiwan: Developing the "TAIDE" (Trustworthy AI Dialogue Engine) to provide a locally-aligned, Mandarin-based chatbot.

Singapore: Developed "SEA-LION" (Southeast Asian Languages in One Network) to improve AI understanding of regional languages.

Brazil: Launched the "Brazilian Artificial Intelligence Plan" (PBIA) with a R$23 billion budget, including upgrading the Santos Dumont supercomputer.

Ukraine: Launching "AI Factory" to build sovereign digital infrastructure for government services and defense.

Kazakhstan: Developing local AI to overcome Western bias and improve Kazakh language representation. 

3.2 Regional and Smaller Initiatives:

Spain: Working on national LLMs covering Spanish, Catalan, Basque, and Galician.

Portugal: Developing "Amalia" through a consortium of universities.

The Netherlands: Developing "GPT-NL" to focus on Dutch language and culture.

Switzerland: Released "Apertus," an open-source model with full transparency on training data. 

<<