Microsoft Research AI is the artificial intelligence division of Microsoft Research, the research subsidiary of Microsoft. It focuses on advancing the state of the art in AI, including machine learning, deep learning, and large language models, and transferring innovations into Microsoft products. The division employs more than 1,000 researchers including computer scientists, physicists, engineers, and mathematicians, among them Turing Award winners, Fields Medal winners, and MacArthur Fellows. Microsoft Research AI is part of the broader Microsoft Research organization, which also includes Microsoft Research NExT and other incubation efforts, all directed by corporate vice president Peter Lee.
The mission of Microsoft Research AI aligns with that of Microsoft Research: to expand the state of the art in research areas, rapidly transfer innovative technologies into Microsoft products, and ensure that Microsoft products have a future. Between 2010 and 2018, Microsoft filed the largest share of AI patents worldwide, accounting for 20% of the 154,000 AI patents filed during that period. Microsoft's annual research spending has been estimated at about $6 billion from 2002 to 2010 and between $10 and $14 billion annually since 2010. Research breakthroughs have been infused into products such as Kinect, Bing, HoloLens, Cortana, Microsoft Translator, LinkedIn, Havok, and Dynamics.
Research Areas
Microsoft Research AI conducts work across a broad spectrum of artificial intelligence topics)Skip: I'll continue.
Microsoft Research AI conducts work across a broad spectrum of artificial intelligence topics, including machine learning, deep learning, and neural networks. Researchers investigate foundational aspects such as optimization algorithms, loss functions, and model pruning, as well as modern techniques like transformers, large language models, and generative AI. Key areas of focus include theory, AI, machine learning, systems and networking, security, privacy, human-computer interaction, and wearable technologies.
The division also contributes to advancements in natural language processing, computer vision, and reinforcement learning. Its work has influenced products across Microsoft, including the Kinect sensor, Bing search, HoloLens mixed-reality headsets, Cortana digital assistant, Microsoft Translator, LinkedIn, Havok physics engine, and Dynamics enterprise software. Collaboration with academic, government, and industry researchers is central to its approach, and Microsoft Research AI maintains strong ties with universities worldwide.
Research Facilities
The Redmond laboratory, founded in 1991 on the Microsoft campus in Washington, houses about 350 researchers and focuses on artificial intelligence, machine learning, systems and networking, security, privacy, human-computer interaction, and wearable technologies. The Cambridge laboratory, established in the United Kingdom in 1997 under Roger Needham, is led by Christopher Bishop and studies machine learning, security, and information retrieval, with close ties to the University of Cambridge.
Microsoft Research Asia, founded in Beijing in November 1998, has grown to include more than 300 researchers and developers alongside about 300 visiting scientists and students, with a satellite office in Shanghai. Its research spans natural user interfaces, multimedia, data-intensive computing, search and online advertising, and natural language processing. The India laboratory, based in Bengaluru and headed by Sriram Rajamani, and the Station Q laboratory at the University of California, Santa Barbara, founded in 2006, explore quantum computing and other topics. Microsoft Research New England, established in 2008 in Cambridge, Massachusetts near the MIT campus, and the New York City laboratory, founded in 2012, focus on interdisciplinary research combining computer science and social science. The Montreal laboratory, established in 2017 following the acquisition of Maluuba, concentrates on natural language processing, deep learning, and reinforcement learning. A Tokyo laboratory was founded in November 2024 to focus on embodied AI, societal AI, and industry innovation.
Former laboratories include the Silicon Valley lab, which operated from August 2001 until 2014 and contributed to research in security, privacy, computer architecture, and internet search.
Key People and Fellowships
Microsoft Research AI includes numerous distinguished researchers, some of whom have received Turing Awards, MacArthur Fellowships, or Dijkstra Prizes. Notable researchers associated with Microsoft Research include Victor Bahl, Anders Hejlsberg, Don Box, Jennifer Tour Chayes, Dave Cutler, David DeWitt, Susan Dumais, Erich Gamma, Jim Gray, Alex Kipman, Butler Lampson, Mark Russinovich, Burton Smith, Jeffrey Snover, Jaime Teevan, and Charles P. Thacker.
The organization sponsors research fellowships for faculty and graduate students in computer science. These fellowships support early-career researchers and have been awarded to many distinguished scientists, including several of the key researchers listed above.
Collaborations and Impact
Microsoft Research AI partners with academic institutions and industry through programs such as joint research initiatives and fellowships. Its laboratories collaborate with universities worldwide, and the Cambridge lab works closely with the University of Cambridge and its Computer Laboratory. The New England and New York City labs, managed by Susan Dumais, foster interdisciplinary research between computer scientists and social scientists.
In the broader AI landscape, Microsoft Research AI competes and cooperates with other major laboratories, including Google DeepMind and OpenAI. Microsoft's investments in generative AI and large language models have positioned it among the leading corporate research organizations in the field, alongside efforts from Amazon Web Services, Google Cloud, and Alibaba Cloud.
The division also engages with academic partners such as MIT CSAIL, Stanford AI Lab, and Carnegie Mellon University. Its researchers regularly collaborate with figures such as Michael I. Jordan and Anima Anandkumar, and the broader field draws on foundational contributions from scholars like Geoffrey Hinton and Yoshua Bengio.
Historical Context and Patents
Since the 1990s, Microsoft Research AI has been a major contributor to the advancement of artificial intelligence. Between 2010 and 2018, Microsoft accounted for the largest share of AI patents filed worldwide, a testament to its research output. This period also saw rapid progress in the field, including the development of key techniques such as multi-head attention, encoder-decoder architectures, and residual networks, all of which have been explored within Microsoft's labs.
Microsoft Research AI has also played a role in the evolution of generative AI and artificial intelligence more broadly. Its researchers have contributed to advancements in machine learning, neural networks, and related disciplines. The division's work on large language models and transformers has informed both academic literature and commercial applications.
Global Presence and Collaborations
The division operates laboratories across multiple continents, including sites in Redmond, Cambridge, Beijing, Bengaluru, New England, New York City, Montreal, and Tokyo. The Montreal laboratory, established in 2017 following Microsoft's acquisition of Maluuba, focuses on natural language processing, deep learning, and reinforcement learning. The Gray Systems Lab, named after Jim Gray, opened in 2008 to research database technologies.
Microsoft Research AI collaborates with external organizations such as the Barcelona Supercomputing Center, INRIA, Carnegie Mellon University, and the São Paulo Research Foundation (FAPESP). It also engages with other AI research groups like Google DeepMind, OpenAI, and academic institutions including MIT CSAIL and Stanford AI Lab. Researchers at Microsoft Research AI have contributed to foundational topics in AI, including neural networks, residual networks, and optimization methods, and have explored areas such as multi-head attention, positional encoding, and beam search.
Notable Contributions and People
Microsoft Research AI has been home to numerous influential researchers. Notable figures include Christopher Bishop, who heads the Cambridge lab and is a recognized authority in pattern recognition and machine learning; Susan Dumais, who manages the New England, New York City, and Montreal labs and has contributed to information retrieval; and Alex Kipman, known for his work on HoloLens and Kinect. Other researchers have advanced fields such as computational social science, distributed computing, and human-computer interaction.
The division has also been involved in the development of AI techniques that underpin modern large language models, including advances in positional encoding, layer normalization, and attention mechanisms. Its contributions to generative AI and neural networks have shaped both Microsoft products and the broader research community.
Research Impact and Future Directions
Microsoft Research AI continues to explore new frontiers, including embodied AI, societal AI, and industry innovation, as reflected in the Tokyo laboratory opened in November 2024. The division's work on reinforcement learning, model pruning, and RLHF (reinforcement learning from human feedback) aims to improve the efficiency, safety, and applicability of AI systems.
With its global network of laboratories and partnerships, Microsoft Research AI remains a central force in the development of artificial intelligence, bridging fundamental research and practical product deployment. Its emphasis on interdisciplinary collaboration ensures ongoing contributions to both academic knowledge and commercial technology.