**Google DeepMind** is a leading British-American AI research laboratory and Alphabet Inc. subsidiary, dedicated to developing safe, responsible artificial intelligence systems that advance scientific discovery and solve complex real-world problems.[1][3] Founded in London in 2010 by Demis Hassabis, Mustafa Suleyman, and Shane Legg, DeepMind pioneered deep reinforcement learning, achieving early breakthroughs like DQN mastering 49 Atari games from raw pixels and WaveNet for realistic text-to-speech in Google Assistant.[1][3] Acquired by Google in 2014 for around $500 million, it gained vast resources to pursue general AI.[1][4] In April 2023, it merged with Google Brain to form Google DeepMind, accelerating efforts amid competition from OpenAI's ChatGPT, under CEO Hassabis.[1][3] Key achievements include AlphaFold (2020), which revolutionized protein folding prediction, releasing 200 million structures for global research; AlphaCode for competitive programming; AlphaDev for faster algorithms; and advances in weather forecasting, nuclear fusion plasma control, robotics (RT-2, PaLM-SayCan), and generative tools like Imagen, Veo, Lyria, and Gemini large language models.[1][3][5] These innovations span healthcare, energy, climate modeling, and speech accessibility via Project Euphonia.[3][4] Today, headquartered in London with global research centers, Google DeepMind drives Alphabet's AI strategy, powering Gemini models and tools like AlphaGenome and WeatherNext.[2][5] In December 2025, it launched a major UK government partnership targeting nuclear fusion, advanced materials (via a 2026 automated lab with Gemini and robotics), AI safety, and societal impacts like labor markets and mental health—bolstered by Google's £5 billion AI investment.[2] It also supports STEM education through scholarships and programs.[7] Ethically focused, DeepMind emphasizes AGI safety while shaping busines