GINC Technology
Artificial Intelligence

Machine Learning & Foundation Models

The capacity to train, tune and field machine learning at the frontier, foundation and general-purpose models, and the model families across language, vision and science built on them.

Capability group
Artificial Intelligence
Domain
Critical Technology · Hard Power
Identifier
factbase.org/cap/AH3D4HX51F
Reference
Machine learning Wikipedia

Papers

680k

2020–2024

Paper growth

+268%

2020–24 on 2015–19

Patent families

506.7k

2020–2024

Filed abroad

15%

families in more than one office

Framework

What it covers

What the capability covers and what it leaves to others, and the indicators it will be measured by.

Bounds

Covers model development as capacity; frontier compute access reads at the control layer through the semiconductor group, and the research workforce through Human Capital, with field position read here.

Lineage

Carried from v0.0 (formerly Machine learning (including neural networks and deep learning)) with the same identifier, absorbing Natural language processing (retired): the field consolidated around general-purpose models, and a taxonomy that held language apart recorded the discipline as it was, not the capability as it is.

Indicators

  • research publication share in machine learningResearch
  • notable frontier models producedPending
  • highly cited AI researchersPending
  • private AI investmentPending
Research

Who publishes it

Papers classified to the capability, by year and by the nation of their authors' institutions, 2020–2024. Read from the v0.0 research topics Machine learning (including neural networks and deep learning) and Natural language processing (including speech and text recognition and analysis).

Papers by year

The five leading nations, papers by year

Share of papers, 2020–2024

Of papers with a known author nation; 37% have none and are left out.

  1. 1China20.6%106,094 papers
  2. 2United States13.5%69,574 papers
  3. 3India13.0%66,831 papers
  4. 4United Kingdom4.0%20,488 papers
  5. 5Germany2.9%14,867 papers
  6. 6Canada2.3%11,935 papers
  7. 7Indonesia2.2%11,342 papers
  8. 8South Korea2.1%10,794 papers
  9. 9Australia2.0%10,411 papers
  10. 10Japan1.9%10,002 papers

Leading institutions, 2020–2024

By fractional count: each paper divided among its authors' institutions.

  1. 1Vellore Institute of Technology University2,175fractional papers
  2. 2SRM Institute of Science and Technology1,828fractional papers
  3. 3Tsinghua University1,736fractional papers
  4. 4Peking University1,638fractional papers
  5. 5Chitkara University1,482fractional papers
  6. 6Amrita Vishwa Vidyapeetham1,398fractional papers
  7. 7University of Electronic Science and Technology of China1,265fractional papers
  8. 8Shanghai Jiao Tong University1,241fractional papers
  9. 9Zhejiang University1,177fractional papers
  10. 10Chinese Academy of Sciences1,150fractional papers
Patents

Who patents it

Patent families by the nation of origin of their applicants and their earliest priority year, from the CPC and IPC classes the capability is read from. 2025 onwards is incomplete: families are published 18 months after priority.

Families by priority year

The five leading origins, families by year

38,535 applicants filed here in 2015–2024; the five largest held 8% of families (HHI 27 of 10,000: unconcentrated).

Share of families by origin, 2020–2024

  1. 1China75.1%380,541 families
  2. 2South Korea8.9%45,029 families
  3. 3United States8.5%43,322 families
  4. 4Japan2.3%11,519 families
  5. 5Germany1.2%6,177 families
  6. 6Taiwan0.5%2,321 families
  7. 7France0.4%1,954 families
  8. 8Canada0.4%1,912 families
  9. 9India0.3%1,638 families
  10. 10United Kingdom0.3%1,618 families

Applicants filing most internationally, 2015–2024

Families filed in more than one office, by applicant as the offices name them.

  1. 1SAMSUNG ELECTRONICS CO LTD4,627international families
  2. 2HUAWEI TECH CO LTD2,552international families
  3. 3MICROSOFT TECHNOLOGY LICENSING LLC2,223international families
  4. 4GOOGLE LLC1,926international families
  5. 5BOSCH GMBH ROBERT1,913international families
  6. 6FUJITSU LTD1,440international families
  7. 7NEC CORP1,429international families
  8. 8IBM1,344international families
  9. 9QUALCOMM INC1,218international families
  10. 10INTEL CORP1,176international families

Classification read

  • G06N3/02 Neural networks: architectures and learning methods · 2.8m documents
  • G06N20/00 Machine learning · 481.4k documents
  • G06N7/01 Probabilistic graphical models · 80.3k documents

Shares inventions with

Capabilities whose classes the same families also fall in, by overlap (Jaccard).

Companies

Who makes it

The companies that research, design, make, supply and operate the technology, the global leaders and the leading firms of every region.

50 companies in this capability, from 21 nations across 8 regions. Sorted by how experts see each company's global relevance in the field, from very high to very low.

  1. Very high5 companies
  2. 1

    United StatesOpenAI

    Developer of the GPT and o-series frontier models and ChatGPT, the most widely used general-purpose foundation model product.

    Very high relevance · Software & platforms · United States, San Francisco · founded 2015

  3. 2

    United StatesAlphabet (Google DeepMind)

    Google DeepMind builds the Gemini frontier model family and scientific models such as AlphaFold, trained on Google's own TPU infrastructure.

    Very high relevance · Software & platforms · United States, Mountain View · founded 1998 · 183k staff

  4. 3

    United StatesAnthropic

    Frontier lab that develops the Claude family of large language models, with a focus on safety research.

    Very high relevance · Software & platforms · United States, San Francisco · founded 2021

  5. 4

    United StatesNVIDIA

    Supplies most of the GPUs and the CUDA software stack used to train foundation models worldwide, and publishes its own Nemotron models.

    Very high relevance · Design · United States, Santa Clara · founded 1993 · 36k staff

  6. 5

    ChinaDeepSeek

    Developer of the open-weight DeepSeek-V3 and R1 models, which reached near-frontier performance at low training cost; backed by the hedge fund High-Flyer.

    Very high relevance · Software & platforms · China, Hangzhou · founded 2023

  7. High10 companies
  8. 6

    United StatesMeta Platforms

    Releases the open-weight Llama model family and runs FAIR and Meta Superintelligence Labs, plus PyTorch, the dominant ML framework.

    High relevance · Software & platforms · United States, Menlo Park · founded 2004 · 74k staff

  9. 7

    United StatesMicrosoft

    Largest investor in and cloud host for OpenAI, operator of Azure AI supercomputers, and developer of its own Phi and MAI models.

    High relevance · Software & platforms · United States, Redmond · founded 1975 · 228k staff

  10. 8

    ChinaAlibaba Group

    Builds the Qwen model family, one of the most widely used open-weight model families globally, and runs China's largest AI cloud.

    High relevance · Operations & infrastructure · China, Hangzhou · founded 1999 · 124k staff

  11. 9

    United StatesxAI

    Develops the Grok frontier models, trained on its Colossus supercomputer in Memphis; merged into SpaceX in 2026.

    High relevance · Software & platforms · United States, Palo Alto · founded 2023

  12. 10

    ChinaByteDance

    Its Seed team builds the Doubao models behind China's most used consumer chatbot, and it is among China's largest buyers of AI compute.

    High relevance · Software & platforms · China, Beijing · founded 2012

  13. 11

    United StatesAmazon (AWS)

    AWS provides model training and hosting (Bedrock, Trainium chips), is Anthropic's main investor and cloud partner, and builds its own Nova models.

    High relevance · Operations & infrastructure · United States, Seattle, Washington · founded 1994 · 1.6m staff

  14. 12

    FranceMistral AI

    Europe's leading frontier model developer, producing open and commercial models (Mistral Large, Magistral, Codestral); ASML is a major investor.

    High relevance · Software & platforms · France, Paris · founded 2023

  15. 13

    ChinaMoonshot AI

    Creator of the Kimi assistant and the open-weight Kimi K2 trillion-parameter mixture-of-experts models.

    High relevance · Software & platforms · China, Beijing · founded 2023

  16. 14

    ChinaZhipu AI (Z.ai)

    A Tsinghua spin-out that builds the GLM model family and listed in Hong Kong in January 2026.

    High relevance · Software & platforms · China, Beijing · founded 2019

  17. 15

    United StatesHugging Face

    Runs the main hub for distributing open models and datasets and maintains the Transformers library; founded by French entrepreneurs.

    High relevance · Software & platforms · United States, New York · founded 2016

  18. Moderate16 companies
  19. 16

    ChinaTencent

    Develops the Hunyuan language, image and 3D model family and uses it across WeChat and Tencent Cloud.

    Moderate relevance · Software & platforms · China, Shenzhen · founded 1998 · 111k staff

  20. 17

    ChinaBaidu

    Develops the ERNIE foundation models and the PaddlePaddle deep-learning framework, and operates Kunlun AI chips.

    Moderate relevance · Software & platforms · China, Beijing · founded 2000 · 36k staff

  21. 18

    ChinaHuawei

    Supplies the Ascend AI accelerators and MindSpore framework that underpin China's domestic training compute, and builds the Pangu models; second-largest international ML patent applicant.

    Moderate relevance · Hardware & devices · China, Shenzhen · founded 1987 · 208k staff

  22. 19

    ChinaMiniMax

    Builds the MiniMax-M text, Hailuo video and speech foundation models; listed in Hong Kong in January 2026.

    Moderate relevance · Software & platforms · China, Shanghai · founded 2021

  23. 20

    CanadaCohere

    Canada's leading foundation model company, offering Command, Embed and Rerank models for enterprise use, plus multilingual Aya research.

    Moderate relevance · Software & platforms · Canada, Toronto · founded 2019

  24. 21

    South KoreaSamsung Electronics

    The largest international patent applicant in machine learning; develops the Samsung Gauss models, and supplies the HBM memory used in AI training.

    Moderate relevance · Components & subsystems · South Korea, Suwon · founded 1969 · 268k staff

  25. 22

    United StatesIBM

    Develops the open Granite enterprise model family and the watsonx platform, and is a top ML patent applicant.

    Moderate relevance · Software & platforms · United States, Armonk · founded 1911 · 270k staff

  26. 23

    South KoreaLG Corp (LG AI Research)

    LG AI Research builds the EXAONE model family, among Korea's leading frontier-class models.

    Moderate relevance · Software & platforms · South Korea, Seoul · founded 1947

  27. 24

    United StatesDatabricks

    Its data and AI platform (including Mosaic AI) is widely used to train, fine-tune and serve models, and it released the DBRX model.

    Moderate relevance · Software & platforms · United States, San Francisco · founded 2013

  28. 25

    United StatesScale AI

    Supplies training data, RLHF labelling and model evaluation to frontier labs and governments; Meta took a 49% stake in 2025.

    Moderate relevance · Professional & data services · United States, San Francisco · founded 2016

  29. 26

    IsraelAI21 Labs

    Israel's main foundation model developer, building the Jamba hybrid Transformer-Mamba models.

    Moderate relevance · Software & platforms · Israel, Tel Aviv · founded 2017

  30. 27

    GermanyBlack Forest Labs

    Founded by the creators of Stable Diffusion, it develops the FLUX image-generation models, among the leading text-to-image models.

    Moderate relevance · Software & platforms · Germany, Freiburg · founded 2024

  31. 28

    United StatesElevenLabs

    Leading developer of speech and voice foundation models; founded by Polish engineers, with major operations in London and Warsaw.

    Moderate relevance · Software & platforms · United States, New York · founded 2022

  32. 29

    United KingdomStability AI

    Developer of the open Stable Diffusion image models and related audio and video generative models.

    Moderate relevance · Software & platforms · United Kingdom, London · founded 2019

  33. 30

    United Arab EmiratesG42

    Abu Dhabi AI group whose Inception unit built the Jais Arabic LLMs; partner with Microsoft and OpenAI in the UAE Stargate campus.

    Moderate relevance · Operations & infrastructure · United Arab Emirates, Abu Dhabi · founded 2018

  34. 31

    United Arab EmiratesTechnology Innovation Institute

    Government-owned research centre under ATRC that develops the open Falcon model family.

    Moderate relevance · Research & IP · United Arab Emirates, Abu Dhabi · founded 2020

  35. Low12 companies
  36. 32

    South KoreaNaver

    Develops HyperCLOVA X, a Korean-centric sovereign large language model used across Naver's services.

    Low relevance · Software & platforms · South Korea, Seongnam · founded 1999

  37. 33

    JapanSakana AI

    Japanese research lab working on nature-inspired model merging and the AI Scientist; co-founded by Transformer co-author Llion Jones.

    Low relevance · Software & platforms · Japan, Tokyo · founded 2023

  38. 34

    JapanPreferred Networks

    Japan's leading deep-learning company; develops the PLaMo models and the MN-Core AI processors.

    Low relevance · Software & platforms · Japan, Tokyo · founded 2014

  39. 35

    JapanFujitsu

    Develops the Takane enterprise model (with Cohere), co-built the Fugaku-LLM, and is a top-ten ML patent applicant.

    Low relevance · Professional & data services · Japan, Tokyo · founded 1935 · 113k staff

  40. 36

    JapanNEC

    Builds the cotomi Japanese large language models and is a top-ten international ML patent applicant.

    Low relevance · Professional & data services · Japan, Tokyo · founded 1899 · 105k staff

  41. 37

    South KoreaUpstage

    Develops the Solar language models and was selected for Korea's national sovereign foundation model programme.

    Low relevance · Software & platforms · South Korea, Seoul · founded 2020

  42. 38

    IndiaSarvam AI

    Selected under the IndiaAI Mission to build India's sovereign foundation model, it releases Indic-language LLM and speech models.

    Low relevance · Software & platforms · India, Bengaluru · founded 2023

  43. 39

    Saudi ArabiaHumain

    Saudi PIF-owned AI company that develops the ALLaM Arabic models and is building national AI data centre capacity.

    Low relevance · Operations & infrastructure · Saudi Arabia, Riyadh · founded 2025

  44. 40

    RussiaYandex

    Develops the YandexGPT models and the Alice assistant, and is Russia's leading ML research employer.

    Low relevance · Software & platforms · Russia, Moscow · founded 1997

  45. 41

    RussiaSberbank

    Russia's majority state-owned bank, which develops the GigaChat language models and the Kandinsky image models.

    Low relevance · Operations & infrastructure · Russia, Moscow · founded 1841

  46. 42

    United KingdomInstaDeep

    Founded in Tunis and still running large African R&D teams, it builds genomics foundation models (Nucleotide Transformer); acquired by BioNTech in 2023.

    Low relevance · Software & platforms · United Kingdom, London · founded 2014

  47. 43

    AustraliaHarrison.ai

    Australia's most prominent foundation model developer, building radiology vision-language models (Harrison.rad) and clinical AI products.

    Low relevance · Software & platforms · Australia, Sydney · founded 2018

  48. Very low7 companies
  49. 44

    TaiwanMediaTek

    MediaTek Research publishes the Breeze traditional-Chinese models, Taiwan's most prominent corporate LLM work, and the company designs edge-AI chips.

    Very low relevance · Design · Taiwan, Hsinchu · founded 1997

  50. 45

    BrazilMaritaca AI

    Brazil's leading LLM developer, producing the Sabiá Portuguese-language model family.

    Very low relevance · Software & platforms · Brazil, Campinas · founded 2022

  51. 46

    VietnamFPT Corporation (FPT Semiconductor)

    Operates Vietnam's NVIDIA-based AI factory and develops Vietnamese-language models through FPT AI.

    Very low relevance · Professional & data services · Vietnam, Hanoi · founded 1988

  52. 47

    South AfricaLelapa AI

    African-language AI lab that released InkubaLM, a small language model for African languages, and the Vulavula speech and NLP platform.

    Very low relevance · Software & platforms · South Africa, Johannesburg · founded 2022

  53. 48

    ThailandSCBX

    Its SCB 10X unit develops Typhoon, Thailand's leading open Thai-language LLM family.

    Very low relevance · Operations & infrastructure · Thailand, Bangkok · founded 1906

  54. 49

    MalaysiaYTL Corporation

    Through YTL AI Labs and YTL AI Cloud it built ILMU, Malaysia's Malay-language LLM, and runs NVIDIA GPU data centres in Johor.

    Very low relevance · Operations & infrastructure · Malaysia, Kuala Lumpur · founded 1955

  55. 50

    NigeriaAwarri Technologies

    Co-developed N-ATLAS, Nigeria's government-backed open multilingual LLM for Yoruba, Hausa and Igbo, with NCAIR (launched September 2025).

    Very low relevance · Software & platforms · Nigeria, Lagos

These companies in the Companies table →