Navigation
Google DeepMind
OrganizationAlso known as: DeepMind, DeepMind Technologies, Google DeepMind
Headcount
2,700
as of Dec 2024
Founded Date
Sep 2010
Key People
4DH
Demis HassabisFounder
Co-founder & CEO
Sep 2010 – present
CEO since founding; became head of Google DeepMind after 2023 Brain merger; 2024 Nobel Prize in Chemistry for AlphaFold
SL
Shane LeggFounder
Co-founder & Chief AGI Scientist
Sep 2010 – present
Leads AGI safety research; PhD on machine super intelligence
MS
Mustafa SuleymanFounder
Co-founder & Head of Applied AI
Sep 2010 – Jan 2022
Left DeepMind in 2022; co-founded Inflection AI, then joined Microsoft as EVP and CEO of Microsoft AI in 2024
PK
Pushmeet Kohli
VP of Research
2017 – present
Leads research across science, safety, and foundational AI
Model Releases
6Dec 2024
Gemini 2.0
Next-generation Gemini with improved agentic capabilities
Feb 2024
Gemini 1.5
Introduced 1M token context window; mixture-of-experts architecture
Dec 2023
Gemini 1.0
Google's first natively multimodal model family (Ultra, Pro, Nano)
Nov 2020
AlphaFold 2
Solved protein structure prediction; accuracy comparable to experimental methods
Dec 2018
AlphaFold
Won CASP13 protein structure prediction competition
Jan 2016
AlphaGo
First AI to defeat a professional Go player; beat Lee Sedol 4-1 in March 2016
All Facts
Financial
Headcount2,700Dec 2024▶
| As Of | Value | Source | Fact ID |
|---|---|---|---|
| Dec 2024 | 2,700 | en.wikipedia.org | f_KtTrbGQynQ |
Organization
Founded DateSep 2010—▶
| As Of | Value | Source | Fact ID |
|---|---|---|---|
| — | Sep 2010 | en.wikipedia.org | f_mEKUPPFYRg |
HeadquartersLondon, UK—▶
| As Of | Value | Source | Fact ID |
|---|---|---|---|
| — | London, UK | deepmind.google | f_pjpNfsuNKw |
Legal StructureSubsidiary of Alphabet Inc.—▶
| As Of | Value | Source | Fact ID |
|---|---|---|---|
| — | Subsidiary of Alphabet Inc. | en.wikipedia.org | f_nq7RXAwqsw |
Safety & Research
Safety Researchers120Jun 2025▶
| As Of | Value | Source | Fact ID |
|---|---|---|---|
| Jun 2025 | 120 | — | f_dANFu1eouA |
People
Founded ByAqcyu3onCA, shane-legg, mustafa-suleyman—▶
| As Of | Value | Source | Fact ID |
|---|---|---|---|
| — | Demis Hassabis, Shane Legg, mustafa-suleyman | en.wikipedia.org | f_AX7AnnLZQQ |
Model Details
Context Window2 millionMay 2024▶
| As Of | Value | Source | Fact ID |
|---|---|---|---|
| May 2024 | 2 million | — | f_4aWyQKbNYA |
General
Websitehttps://deepmind.google/—▶
| As Of | Value | Source | Fact ID |
|---|---|---|---|
| — | https://deepmind.google/ | wikidata.org | 2Y4spe8DsA |
Research Areas
4| Name | Description | Team Size | Started | Key Publication | Notes |
|---|---|---|---|---|---|
| Game-Playing AI | Reinforcement learning for complex games including Go, chess, StarCraft, and Diplomacy | — | Jan 2013 | nature.com | AlphaGo (2016), AlphaZero (2017), AlphaStar (2019) |
| Protein Structure Prediction | Deep learning for predicting 3D protein structures from amino acid sequences | — | Jan 2016 | nature.com | AlphaFold 2 solved the 50-year protein folding problem; Demis Hassabis received 2024 Nobel Prize in Chemistry |
| Neuroscience-Inspired AI | Using insights from neuroscience to improve AI architectures and learning algorithms | — | Sep 2010 | — | Core research area since founding; memory systems, attention, and reward-based learning |
| AI Safety | Responsible AI development including alignment, robustness, and societal impact | — | Jan 2017 | deepmind.google | Frontier Safety Framework published 2024; safety team includes alignment, interpretability, and ethics researchers |
Safety Milestones
3| Name | Date | Type | Description | Source | Notes |
|---|---|---|---|---|---|
| Specification Gaming Research | Apr 2020 | research-paper | Catalogued examples of AI systems exploiting reward misspecification | deepmind.google | Influential taxonomy of reward hacking failures |
| Frontier Safety Framework | May 2024 | policy-update | Framework for evaluating and mitigating risks from frontier AI models | deepmind.google | Defines Critical Capability Levels (CCLs) analogous to Anthropic's ASLs |
| Dangerous Capability Evaluations | Oct 2023 | safety-eval | Systematic evaluations for dangerous capabilities in frontier models | arxiv.org | Published jointly with DeepMind safety team; covers bio, cyber, autonomy, and persuasion risks |
Strategic Partnerships
3| Type | Date | Investment Amount | Compute Commitment | Source | Notes | Partner |
|---|---|---|---|---|---|---|
| acquisition | Jan 2014 | 500,000,000 | — | en.wikipedia.org | Acquired by Google for approximately $500M; conditions included creation of AI ethics board | Google (Alphabet) |
| merger | Apr 2023 | — | — | blog.google | DeepMind merged with Google Brain to form Google DeepMind; Jeff Dean became Chief Scientist of Google DeepMind and Google Research | Google Brain |
| spin-off | Nov 2021 | — | — | isomorphiclabs.com | Drug discovery spin-off from DeepMind, led by Demis Hassabis; leverages AlphaFold technology | Isomorphic Labs |
▶Internal Metadata
| ID | A4XoubikkQ |
| Stable ID | A4XoubikkQ |
| Numeric ID | E98 |
| Type | organization |
| YAML Source | packages/kb/data/things/A4XoubikkQ.yaml |
| Facts | 8 structured (10 total) |
| Records | 20 in 5 collections |