Microsoft Research
Microsoft Research advances science and technology to benefit humanity. It conducts fundamental science and technology research across the computing and AI stack, pursuing breakthroughs that accelerate discovery across sciences and create foundations for others to build upon.
At a Glance
- Microsoft product and engineering teams
- Academic researchers, universities and students
- Government and public-sector research organizations
- AI developers and open-source communities
- +2 more
AI Tools by Microsoft Research
(6)
MarkItDown
Python Files to Markdown Converter
ONNX Runtime
Cross Platform ML Inference Engine
Skill Recorder
Work Session to AI Skill Recorder
Ontology Playground
Ontology Design and Explorer Tool
Flint Chart
AI Chart Spec Language for Agents
AutoGen
Open Source Multi-Agent Framework
Discussions
No discussions yet
Be the first to start a discussion about Microsoft Research
Latest News
GigaPath-Flash and GigaTIME-Flash introduced as efficient open pathology foundation models for population-scale discovery.
MindTopo introduced as a benchmark for multimodal models' topological reasoning and planning.
CARE-X introduced as a research radiology vision-language model with auxiliary supervision, reward-aligned learning and tool-augmented measurement.
Orchard introduced as an open framework for scalable agentic AI, with Orchard Env and domain-specific recipes.
Products & Services
Lightweight Python utility that converts PDFs, Office documents and other file types into Markdown while preserving useful structure such as headings, lists, tables and links; intended for LLM and text-analysis pipelines. It is MIT licensed and supports optional plugins and Azure Content Understanding integration.
Open-source human-centered web-agent research prototype that collaborates with users on browser tasks, including web browsing, form filling, file understanding, and Python or shell execution; integrated with Azure AI Foundry models and agents.
Open-weight, 7-billion-parameter agentic small language model for computer use. It visually perceives webpages and operates mouse and keyboard interfaces, with releases through Microsoft Foundry, Hugging Face and Copilot+ PC/Windows AI Toolkit paths.
Open-weight pathology foundation models designed to reduce the compute cost of whole-slide representation learning and H&E-to-multiplex-immunofluorescence translation for population-scale cancer research; released under Apache 2.0.
Market Position
Microsoft Research is a large, globally distributed corporate research lab that combines an open academic research model with a direct path to technology transfer into Microsoft's cloud, software and AI products. In corporate AI and computer-science research it operates in the same broad competitive landscape as Google DeepMind, Meta AI, IBM Research and OpenAI, while differentiating through Microsoft's Azure/Foundry ecosystem, global labs and breadth across computing and scientific disciplines.
Leadership
Founders
Richard Rashid
Computer scientist and Microsoft executive who led Microsoft Research for many years as its senior vice president; he previously held an academic research career in computer science.
Bill Gates
Microsoft co-founder and chairman at the time Microsoft Research was created; he championed an open, collaborative environment for long-horizon computing research.
Nathan Myhrvold
Former Microsoft chief technology officer and physicist; he helped establish the research arm's long-range, fundamental-research orientation before later founding Intellectual Ventures.
Executive Team
Peter Lee
President, Microsoft Science; previously leader of the worldwide laboratories of Microsoft Research
Previously established a technology office at DARPA focused on machine learning, data science and computational social science; before joining Microsoft in 2010, he was a Carnegie Mellon University professor and head of its computer science department.
Lili Cheng
Corporate Vice President and Managing Director, Microsoft Research Americas
Previously Corporate Vice President and CTO of Microsoft's Business Applications Group; earlier founded Microsoft Research's Social Computing Research group, led FUSE Labs, and worked at Apple Computer's Advanced Technology Group on QuickTime Conferencing and QuickTime VR.
Board of Directors
Founding Story
Microsoft created its own computer-science research organization in 1991 to pursue both basic and applied research beyond current product cycles. The initial vision was an open academic-style environment where researchers could tackle difficult computing problems independently, collaborate with universities and industry, and transfer breakthroughs into Microsoft products and the wider technology ecosystem.
Business Model
Revenue Model
Microsoft Research is an internally funded corporate research organization rather than a separately marketed subscription business. Its economic impact comes through technology transfer into Microsoft products and services, intellectual-property licensing and open-source/open-weight research releases; the 2006 Microsoft account also described an IP Ventures program that licensed technology to startups.
Target Markets
- Microsoft product and engineering teams
- Academic researchers, universities and students
- Government and public-sector research organizations
- AI developers and open-source communities
- Healthcare, biomedical and pharmaceutical researchers
- Enterprises exploring AI, agents and scientific computing
- AI and machine-learning research and model development
- LLM grounding, claim verification and document ingestion
- Agentic software engineering, web navigation and personal-assistant workflows
- Computer-use automation with user oversight
- Digital pathology, cancer research and population-scale biomedical discovery
- Scientific discovery in materials, chemistry, climate and drug development
- Microsoft product groups and customers benefiting from transferred research innovations
- University of Washington and Providence
- Narayana Health
- Nissan Motor Corporation