Reference
AI organisations
A directory of the organisations that matter to this record — the labs racing to build frontier models, the safety and policy groups arguing over how they should be governed, the evaluators keeping score, the hardware makers supplying the compute, and the companies putting it all to work. Each profile links to the organisation’s full timeline in the record. The focus is on AI-first organisations rather than firms that merely use AI as a feature; the big-tech backers and media rightsholders that recur as investors and litigants trail at the end.
176 organisations · 15 categories
Frontier & foundation-model labs
The labs training general-purpose frontier models — the ones the field measures itself against.
OpenAI
The lab that built the GPT models and ChatGPT and set off the commercial AI race; still among the most-used and most-valuable AI companies in the world.
Sam Altman, Greg Brockman~$500bn valuation (2025); backed by Microsoft and SoftBank
Anthropic
A frontier lab building the Claude family of models, founded on the argument that the companies racing to build powerful AI should be the ones most invested in keeping it controllable.
Dario Amodei, Daniela Amodei~$965bn valuation (2026 Series H); backed by Amazon and Google
Google DeepMind
Alphabet's AI division, formed in 2023 by merging the London lab DeepMind with Google Brain; builds the Gemini models and the AlphaFold line of scientific systems.
Meta AI
Meta's AI effort, from the FAIR research lab to the 2025 Meta Superintelligence Labs division; long the standard-bearer for open-weight models before shifting toward closed releases.
Mistral AI
Europe's leading independent lab, known for shipping capable open-weight models under permissive licences alongside its Le Chat assistant.
Arthur Mensch, Timothée Lacroix, Guillaume Lample€1.7bn round led by ASML (2025)
Cohere
An enterprise-focused lab building large language models for businesses and governments rather than the consumer market, with an emphasis on private deployment.
Aidan Gomez, Nick Frosst, Ivan Zhang$500m Series D at $6.8bn valuation (2025)
Thinking Machines Lab
Mira Murati's research-and-product company, built from a largely ex-OpenAI team and pitched around adapting AI to specific human needs with more open research than the largest labs.
Mira Murati, John Schulman$2bn seed at $12bn valuation (2025), reported the largest first round on record
AI21 Labs
An Israeli lab, among the first outside OpenAI to sell a large language model via API; later shifted to its hybrid Jamba architecture and enterprise reliability tools.
Ori Goshen, Yoav Shoham, Amnon ShashuaReported $300m Series D (2025) later said by Calcalist never to have closed
Safe Superintelligence
Ilya Sutskever's research company, set up around the single stated goal of a "safe superintelligence" with no plan to sell products along the way.
Ilya Sutskever, Daniel Levy~$32bn valuation (2025); strategic compute partnership with Nvidia (2026)
Reflection AI
A lab that repositioned in 2025 toward open-weight frontier models, styling itself an American counterpart to China's DeepSeek.
Misha Laskin, Ioannis Antonoglou$2bn at $8bn valuation (Oct 2025), led by Nvidia
Aleph Alpha
Sovereign, enterprise-focused European LLMs; pivoted in 2024 from training its own frontier models to a platform integrating third-party ones.
Jonas Andrulis, Reto Spörri, Ilhan Scheer$500m Series B (2023), backed by Schwarz Group, SAP and Bosch
Liquid AI
Efficient, on-device foundation models based on "liquid neural networks", a non-transformer design spun out of MIT.
Ramin Hasani, Daniela Rus$250m Series A at ~$2.35bn valuation (Dec 2024), led by AMD Ventures
Poolside
Foundation models for software engineering, sold with on-premises deployment for regulated industries.
Jason Warner, Eiso KantNvidia to invest up to $1bn at ~$12bn valuation (Oct 2025)
Sakana AI
Evolutionary and model-merging methods rather than ever-larger training runs, with models tuned for Japan.
David Ha, Llion Jones, Ren Ito$135m Series B at $2.65bn valuation (Nov 2025)
Inflection AI
Founded to build the Pi personal-AI companion; in 2024 Microsoft hired its founders and most of its staff and licensed its technology, leaving a smaller, nominally independent enterprise business.
Sean White, Reid Hoffman>$1.3bn raised (2023); backers included Microsoft, Nvidia and Bill Gates
Imbue
Reasoning-focused AI agents and developer tools, framed around amplifying rather than replacing the user.
Kanjun Qiu, Josh Albrecht$200m+ Series B at ~$1bn valuation (2023), backed by Nvidia and Astera Institute
Chinese AI labs
China’s model builders, many shipping open weights under permissive licences.
DeepSeek
Open-weight large language and reasoning models, spun out of the High-Flyer quantitative hedge fund.
Alibaba / Qwen
The Qwen family of open-weight and frontier models, developed by the company's cloud division.
Zhipu AI
The open-weight GLM family of models, spun out of Tsinghua University; now listed as Z.ai.
Moonshot AI
The Kimi assistant and open-weight Kimi models, known first for long-context handling and later for coding and agentic tasks.
ByteDance
The Doubao assistant and the Seed research group's model family, from the company behind TikTok and Douyin.
Tencent
The Hunyuan family of language, image and video models, from the company behind WeChat.
Baidu
The ERNIE family of models and the ERNIE Bot assistant, from China's dominant search company.
MiniMax
Open-weight foundation models and the Hailuo platform for video and audio generation.
01.AI
Formerly a builder of open-weight Yi foundation models; since 2025 refocused on fine-tuning others' open models into enterprise software after stepping back from pretraining from scratch.
Shanghai AI Laboratory
A government-backed research institute developing the open-source InternLM models and OpenGVLab vision projects.
iFlytek
Speech technology and the Spark (Xinghuo) large-model family, from a long-established, partly state-owned Chinese AI firm.
StepFun
Multimodal foundation models — the Step series, spanning language, vision and audio.
Jiang DaxinTencent-backed
Beijing Academy of Artificial Intelligence
A non-profit research institute known for the Wu Dao models and for open datasets and tools for Chinese AI research.
Open source & ecosystem
The hubs, datasets and open-model builders the wider field depends on.
Hugging Face
The main hub for sharing open-weight models and datasets, and the maker of open-source libraries such as Transformers.
Clément Delangue, Julien Chaumond, Thomas Wolf~$4.5bn valuation (2023 Series D led by Salesforce Ventures)
Allen Institute for AI
A non-profit research institute known for fully open language models (OLMo) that publish training data, code and checkpoints, not just final weights.
EleutherAI
A research collective turned non-profit that pioneered openly released language models (GPT-Neo, GPT-J) and open training corpora (The Pile).
Together AI
A cloud platform for running and fine-tuning open-weight models across a large catalogue, positioned as infrastructure for the open ecosystem.
Vipul Ved Prakash, Tri Dao~$3.3bn valuation (2025 Series B); ~$1.33bn raised, backers include Nvidia and General Catalyst
LAION
A German non-profit that builds and releases large open image-text datasets used to train open generative-image models.
Nous Research
An open-weight model group known for its Hermes fine-tunes and for decentralised-training work such as the Psyche network.
Karan Malhotra, Jeffrey Quesnelle~$1bn valuation (2025 Series A led by Paradigm); reported in mid-2026 to be raising more at ~$1.5bn
OpenRouter
A unified API and marketplace that routes requests across many LLM providers through a single endpoint.
Alex Atallah~$1.3bn valuation (Series B led by CapitalG); ~$174m raised
Ollama
A popular open-source tool for running open-weight models locally on a personal computer.
Jeffrey Morgan, Michael Chiang~$88m raised; $65m Series B (2026) led by Theory Ventures
Arcee AI
Small, domain-specific language models and model-merging tools for enterprises.
Mark McQuade, Jacob Solawetz$24m Series A (2024), led by Emergence Capital
Motif Technologies
Builds open-weight models for South Korea's sovereign-AI programme, trained without foreign model weights.
Generative media
Image, video, audio, music and voice generation.
Stability AI
The company behind Stable Diffusion, the first capable image generator released with openly downloadable weights, and a standard-bearer for the open-weights camp in generative media.
Prem AkkarajuRaised at ~$1bn valuation (2022); restructured in 2024 under new leadership after Emad Mostaque's departure
Midjourney
A small, self-funded lab whose Discord-based image generator, known for a painterly house style, has been among the most widely used since it opened in 2022.
David HolzSelf-funded; the company says it is profitable without having raised venture capital
Suno
An AI music company whose app turns a short text prompt into a complete song — vocals, lyrics and instrumentation — in about a minute.
Mikey ShulmanRaised $125m (2024); reported ~$2.45bn valuation in a 2025 round
Runway
A New York company making generative video and image tools, and an early mover in putting text-to-video generation in front of the public.
Cristóbal Valenzuela~$3bn valuation (2025 Series D)
ElevenLabs
A voice-AI company whose tools generate realistic synthetic speech and can clone a voice from a short sample, used for narration, dubbing and conversational agents.
Mati Staniszewski, Piotr Dąbkowski$3.3bn valuation (Jan 2025 Series C); reportedly raised further later in 2025
Black Forest Labs
Maker of the open-weights FLUX family of image-generation models, founded by researchers from the original Stable Diffusion team.
Robin Rombach, Andreas Blattmann, Patrick Esser$300m round (Dec 2025) at ~$3.25bn valuation; backers include Andreessen Horowitz and Nvidia
Udio
An AI music-generation platform, built by former Google DeepMind researchers, that produces songs from text prompts.
David Ding, Andrew SanchezSeed round led by Andreessen Horowitz (2024)
fal
A cloud platform serving a large catalogue of third-party generative-media models via fast API, now also shipping its own tuned models.
Ideogram
Text-to-image generation, noted for rendering legible text within generated images.
Mohammad Norouzi, Jonathan Ho$80m Series A (2024), led by Andreessen Horowitz; ~$96.5m raised in total
Pika
Consumer-facing tools for generating and editing short video clips from text and images.
Demi Guo, Chenlin MengRaised ~$135m to date; backers include Lightspeed Venture Partners
Cartesia
Low-latency, real-time voice generation built on state-space model architectures.
Karan Goel, Albert Gu$100m round (backers include Nvidia and Kleiner Perkins); $64m Series A (2024)
Hume AI
Emotionally expressive speech models and an "empathic" conversational voice interface.
Alan Cowen$50m Series B (2024), led by EQT Ventures
Krea AI
A real-time creative canvas bundling the company's own and third-party image and video models.
Victor Perez, Diego Rodriguez~$500m valuation (2025 Series B led by Bain Capital Ventures); ~$83m raised in total
Agents & applied AI
AI-native products and agentic companies built around models rather than training frontier ones.
Perplexity
A search company built around an "answer engine" that retrieves web pages and uses a language model to write a directly cited answer rather than returning a list of links.
Aravind Srinivas~$20bn valuation (2025)
Cursor / Anysphere
An AI-assisted code editor, made by Anysphere and built as a fork of VS Code, that became one of the fastest-growing developer tools by revenue.
Michael Truell~$29.3bn valuation (2025); agreed 2026 acquisition by SpaceX
Harvey
AI tools for legal and professional-services work — contract review, due diligence, research and drafting — sold mainly to large law firms.
Winston Weinberg, Gabriel Pereyra~$11bn valuation (2026); backed by the OpenAI Startup Fund, Sequoia and Kleiner Perkins
Character.AI
A consumer platform that lets people chat with and design AI "characters" with distinct personalities, widely used among younger users.
Lovable
Natural-language "vibe-coding" that generates and deploys web apps; among Europe's highest-valued AI startups.
Replit
A browser-based IDE, plus an AI coding Agent that builds working software from natural-language descriptions.
Amjad Masad~$9bn valuation (2026 Series D)
Glean
Enterprise search and an AI work assistant that indexes a company's internal tools and answers questions across them.
Arvind Jain~$7.2bn valuation (2025 Series F)
Sierra
AI agents that handle customer-service conversations by chat and voice for businesses.
Bret Taylor, Clay Bavor~$15.8bn valuation (2026); backed by Sequoia, Benchmark and Greenoaks
WindsurfAcquired
An AI-assisted code editor, formerly Codeium; its leadership was hired by Google in 2025 and the rest of the company acquired by Cognition.
Varun Mohan, Douglas ChenSplit in 2025: ~$2.4bn Google licence-and-hire; ~$250m Cognition purchase
Decagon
AI agents that handle customer-support conversations across chat, email and voice for enterprises.
Jesse Zhang, Ashwin Sreenivas~$1.5bn valuation (2025); reported ~$4.5bn (2026)
Robotics & embodied AI
Humanoids, self-driving and the effort to put models into the physical world.
Waymo
Autonomous driving; runs a commercial driverless robotaxi service across several US cities.
Tekedra Mawakana, Dmitri DolgovAlphabet subsidiary; valued ~$126bn in a 2026 external financing round
Figure AI
General-purpose humanoid robots aimed at manual labour in industry and, eventually, the home.
Brett Adcock~$39bn valuation (Series C, Sept 2025); backers include Nvidia, Microsoft, Bezos Expeditions
Unitree Robotics
Maker of quadruped and humanoid robots, among the most prominent names in China's humanoid push.
Wang Xingxing~$53bn market value at its August 2026 Shanghai listing
Physical Intelligence
General-purpose foundation models ("π") intended to control many kinds of robot from natural-language instructions.
Karol Hausman, Sergey Levine, Chelsea Finn~$5.6bn valuation (Series B, Nov 2025); backers include Jeff Bezos, OpenAI
World Labs
"Large world models" for spatial intelligence — AI that perceives and generates 3D environments.
Fei-Fei Li~$1.23bn raised across two rounds (2024-26); backers include Nvidia, AMD, Autodesk
1X Technologies
Humanoid robots for the home; its consumer robot NEO opened for pre-order in 2025.
Bernt BørnichReported to be raising ~$1bn at ~$10bn valuation (2025); backed by the OpenAI Startup Fund
Apptronik
General-purpose humanoid robots (Apollo) for industrial and logistics work.
Jeff Cardenas~$5bn valuation (2026); Series A co-led by Google and B Capital
Skild AI
A general-purpose "robot brain" foundation model designed to run across many different robot bodies.
Deepak Pathak, Abhinav Gupta~$14bn valuation (Series C, Jan 2026, led by SoftBank)
Wayve
End-to-end learned self-driving software, pitched as an "embodied AI" layer that carmakers can license.
Alex Kendall~$8.6bn valuation (Series D, 2026); backers include SoftBank, Nvidia, Microsoft, Uber
X-Humanoid
State-backed Beijing developer of the Tiangong humanoid robots.
Nuro
Autonomous driving; began with delivery vehicles and now licenses its "Nuro Driver" system.
Dave Ferguson, Jiajun Zhu~$6bn valuation (2025 round), down from a $8.6bn peak (2021)
Data, training & services
The human-labelled data, evaluation and RLHF layer that frontier training runs on.
Scale AI
Data company supplying much of the human-labelled and human-evaluated training data used by frontier labs; expanded into benchmarks and evaluation tools.
Jason DroegeMeta paid $14.3bn for a 49% non-voting stake (2025)
Databricks
Data and analytics platform, built by the creators of Apache Spark, that helps businesses train and fine-tune their own models on their own data.
Ali GhodsiReported ~$134bn valuation (2025)
Mercor
Marketplace matching frontier labs with vetted domain experts — doctors, lawyers, scientists, engineers — to produce and evaluate model-training data.
Brendan Foody, Adarsh Hiremath, Surya Midha$350m Series C at a $10bn valuation (Oct 2025); reported in 2026 in talks at ~$20bn
Surge AI
Human-labelled data and reinforcement-learning-from-human-feedback (RLHF) services for frontier labs, supplied by a large network of human annotators.
Edwin ChenBootstrapped (no VC); valued at ~$20bn in a 2025 secondary sale, per Forbes
Snorkel AI
Enterprise platform for "programmatic" data labelling — labelling training data with rules and weak supervision — plus expert-curated datasets and evaluators for frontier models.
Alex Ratner$100m Series D at a reported ~$1.3bn valuation (2025)
Turing
Coding, reasoning and agentic training data and human expertise for frontier labs, plus AI systems built for large enterprises; grew out of a remote-engineering marketplace.
Jonathan Siddharth, Vijay Krishnan$111m Series E at a $2.2bn valuation (2025)
Compute, chips & infrastructure
Accelerators, wafer-scale and neocloud capacity — the hardware the boom runs on.
NVIDIA
The chip designer whose GPUs became the default hardware for training and running large AI models, making it the biggest financial beneficiary of the boom.
Jensen HuangPublicly listed (Nasdaq: NVDA); first company to reach a $5tn market value (2025)
CoreWeave
One of the largest "neocloud" providers, renting out Nvidia GPU capacity to AI labs and other firms that need large amounts of compute but do not want to build their own data centres.
Michael IntratorPublicly listed (Nasdaq: CRWV) after a March 2025 IPO; backed by Nvidia
Huawei
Chinese electronics giant whose Ascend accelerators are the main domestically made alternative to Nvidia GPUs for AI developers inside China under US export controls.
ASML
The sole maker of extreme-ultraviolet lithography machines needed to manufacture the most advanced AI chips, sitting upstream of every leading-edge foundry.
Christophe FouquetPublicly listed (Nasdaq/Euronext: ASML)
Cerebras
AI-chip maker known for wafer-scale processors the size of a dinner plate, pitched as an alternative to Nvidia GPUs particularly for fast model inference.
Andrew Feldman~$56bn valuation at its 2026 Nasdaq IPO
Qualcomm
Mobile and on-device AI chip leader (Snapdragon) that in 2025 entered the data-centre market with the AI200 and AI250 inference accelerators.
Cristiano AmonPublicly listed (Nasdaq: QCOM)
Crusoe
An energy-first AI infrastructure company that builds and operates data centres and a GPU cloud, siting compute where power is cheap.
Chase Lochmiller, Cully Cavness$1.375bn Series E (Oct 2025) at a reported valuation above $10bn
Etched
Building "Sohu", a chip specialised for the transformer architecture, betting that a fixed-function design beats general-purpose GPUs on inference speed and cost.
Gavin Uberti, Chris ZhuOver $1bn raised; reported ~$5bn valuation (late 2025)
Groq
Designs "Language Processing Unit" chips aimed at fast, low-cost AI inference, an alternative to Nvidia GPUs for running rather than training models.
Jonathan Ross~$6.9bn valuation (2025); Nvidia struck a reported ~$20bn licensing deal in Dec 2025
Lambda
A GPU cloud provider renting Nvidia capacity for training and running AI models, plus AI workstations and software.
Stephen Balaban, Michael Balaban$480m Series D (Feb 2025) at a reported ~$2.5bn valuation; later raised $1.5bn+ (Series E)
Nebius
Builds and rents GPU cloud capacity for AI, spun out of the European assets of the Russian group Yandex and now listed on Nasdaq.
Arkady Volozh$700m private placement (Dec 2024) from investors including Nvidia and Accel; listed on Nasdaq (NBIS)
SambaNova Systems
Builds AI chips and systems on a "reconfigurable dataflow" architecture, an alternative to GPU-based designs focused on running large models efficiently.
Rodrigo Liang~$11bn valuation in a $1bn Series F (2026)
Tenstorrent
Builds AI accelerators and CPU cores on the open RISC-V architecture, sold both as hardware and as licensable IP, an open alternative to proprietary GPU designs.
Jim Keller$693m Series D (Dec 2024) at a reported ~$2.6bn valuation; backers include Samsung and Hyundai
TaalasAcquired
Hardwires trained model weights into dedicated silicon for low-cost, high-speed inference; acquired by AMD in 2026.
Safety & alignment research
Independent research on control, interpretability and dangerous-capability evaluation.
METR
A nonprofit that runs independent pre-deployment evaluations of frontier models for dangerous autonomous capabilities, and is known for its "time horizon" measure of AI progress.
Center for AI Safety
A nonprofit combining technical AI-safety research, benchmarks and field-building with public advocacy; author of the 2023 Statement on AI Risk.
Redwood Research
A small nonprofit that developed the "AI control" agenda — safeguards designed to hold even if a deployed model is actively trying to subvert them — and runs empirical safety work with larger labs.
Apollo Research
A nonprofit that evaluates frontier models for deceptive and "scheming" behaviour and runs pre-deployment safety tests for AI labs.
MIRI
One of the oldest AI-risk organisations, focused on the theory of aligning advanced AI and, more recently, on advocacy for slowing frontier development.
Alignment Research Center
A nonprofit doing theoretical AI alignment research — eliciting latent knowledge and heuristic arguments — whose evaluations team spun out to become METR.
Goodfire
A mechanistic-interpretability startup building tools to inspect, debug and steer the internals of neural networks, sold as a commercial platform.
Eric Ho, Tom McGrath$50m Series A (2025), led by Menlo Ventures, with Anthropic
Gray Swan AI
Red-teaming of frontier models and jailbreak-resistance defences (Cygnal, Shade), plus a crowdsourced adversarial-testing arena.
Matt Fredrikson, Zico Kolter, Andy Zou$40m Series A (2025), led by Madrona and Wing Venture Capital
Palisade Research
Empirical research demonstrating the offensive and control-resistant capabilities of frontier AI agents, including shutdown-resistance and autonomous-hacking tests.
Transluce
A nonprofit building open tools to understand and oversee AI systems, including automated "investigator agents" and interpretability methods.
Conjecture
An alignment-focused startup pursuing "cognitive emulation" — bounded, auditable AI systems it argues are safer than scaling ever-larger models.
Eleos AI Research
Research on AI moral patienthood and model welfare — whether and when AI systems might be candidates for moral concern, and how to evaluate them.
FAR AI
An AI-safety research and field-building nonprofit that does technical work on frontier-model robustness and runs a Berkeley research hub, grants and conferences.
Adam Gleave, Karl Berzins>$30m from multiple funders (announced Jan 2026)
Haize Labs
Automated red-teaming and large-scale adversarial stress-testing of AI models.
Leonard Tang, Richard Liu, Steve Li$12.5m seed (2024), led by General Catalyst, reported at a ~$100m valuation
Timaeus
Developmental interpretability and singular learning theory — understanding how structure emerges in neural networks during training.
Evaluation & forecasting
Benchmarks, leaderboards and the analysts tracking how fast AI is moving.
Epoch AI
Nonprofit research institute that measures and forecasts the trajectory of AI — compute, data, hardware and investment — and built the FrontierMath benchmark.
LMArena
Public leaderboard, originally the Berkeley-linked Chatbot Arena, that ranks AI models by anonymous head-to-head human votes rather than fixed test sets.
Anastasios Angelopoulos, Wei-Lin Chiang, Ion Stoica$100m seed led by a16z and UC Investments (2025)
Artificial Analysis
Independent benchmarking of AI models and inference providers on intelligence, speed and price, published as public leaderboards and a composite Intelligence Index.
Micah Hill-Smith, George CameronRaised ~$2.6m seed (2024)
Vals AI
Independent, third-party evaluation of language models on domain-specific professional tasks in law, tax, finance and other regulated fields.
Rayan Krishnan, Langston Nashold~$5m raised (2024); backed by Sequoia Capital and Bloomberg Beta
Policy, governance & think tanks
Non-profits, institutes and advocacy shaping how AI is governed.
Future of Life Institute
A non-profit working to reduce catastrophic and existential risks from advanced technology; author of the 2023 open letter calling for a pause on giant AI experiments.
RAND
A long-established US non-profit research institution whose AI work spans frontier-model governance, national security, autonomous weapons and the economics of automation.
CSET
A Georgetown research centre providing non-partisan analysis of AI and national security — talent pipelines, semiconductors, China's AI ecosystem and compute.
Dewey MurdickLaunched with a $55m grant from Open Philanthropy (2019)
AI Policy Institute
A US think tank that polls the public on AI and advocates government action on catastrophic risk, widely cited for its opinion surveys.
Daniel ColsonSays it was seeded by anonymous tech and finance donors; reporting also links funding to Open Philanthropy
Americans for Responsible Innovation
US non-profit advocating bipartisan federal AI policy that it says balances public-interest safeguards with maintaining American competitiveness.
Brad Carson, Eric GastfriendDonor-funded; backers reported to include the Omidyar Network and Inclusive Abundance Action
Ada Lovelace Institute
An independent, Nuffield-funded research body examining how data and AI affect society, known for its work on biometrics, facial recognition and the regulation of AI in the public sector.
Gaia MarcusFounded and funded by the Nuffield Foundation (est. with an initial £5m commitment)
AI Futures Project
A small research non-profit forecasting AI's trajectory; author of the widely discussed "AI 2027" scenario.
AI Now Institute
A critical AI-policy research institute focused on the concentration of power in the tech industry, algorithmic bias, labour and surveillance.
CNAS
A bipartisan US national-security think tank whose technology work covers military AI, autonomous weapons, escalation risk and export controls.
Encode
Youth-led AI-policy advocacy group (formerly Encode Justice) that campaigns for AI regulation and intervened in the OpenAI restructuring and Musk litigation.
GovAI
An Oxford-linked research non-profit studying the governance of advanced AI — frontier-model regulation, corporate governance and security implications.
Leading the Future
Pro-AI US super PAC network backing candidates favourable to AI development and opposing those seeking tighter restrictions.
Reported at $100m+ (2025); backers include Andreessen Horowitz, Greg Brockman and Joe Lonsdale
Partnership on AI
A multi-stakeholder non-profit consortium of companies, academics and civil-society groups developing best practices and shared guidance for responsible AI.
PauseAI
International grassroots movement campaigning for a pause on frontier AI development until safety can be demonstrated.
Wadhwani AI Center
The AI programme of the CSIS think tank, analysing AI through a national-security lens — export controls, US–China competition, defence adoption and governance.
Aalok MehtaLaunched with a $5m commitment from CSIS trustee Romesh Wadhwani (2023)
ControlAI
A UK-and-US non-profit campaigning for binding international limits on the development of artificial superintelligence.
Andrea Miotti, Connor LeahyUK non-profit and US 501(c)(4); does not publish donors, funding sources not publicly known
National Republican Senatorial Committee
The campaign arm of US Senate Republicans; notable in the AI record for airing an attack ad built on a full AI recreation of an opposing candidate.
AI Policy Network
A Washington non-profit lobbying Congress on advanced-AI and national-security policy, pressing for federal rules on the most powerful systems.
Mark BeallRegistered as a 501(c)(4), which need not disclose donors; funding sources not publicly known
Alliance for Secure AI
A US non-profit campaigning for bipartisan guardrails on advanced AI through public-awareness and advocacy campaigns.
Brendan SteinhauserRuns a 501(c)(3) and an affiliated 501(c)(4); donors not disclosed, backers not publicly known
Alliance for the Future
US non-profit advocating a pro-innovation, anti-restriction stance on AI policy and opposing what it terms AI "doomer" positions.
Center for AI Policy
Washington non-profit advocating binding US federal regulation of advanced AI, drafting model safety legislation and lobbying Congress.
Forethought
Macrostrategy research on preparing for advanced AI — power concentration, the intelligence explosion, and steering the transition rather than only avoiding catastrophe.
Collective Action for Existential Safety
Grassroots coordination for existential safety, with a public catalogue of actions individuals, organisations and governments can take.
James Norris
Governments & standards bodies
State institutes, regulators and multilateral bodies setting the rules.
US Government
Sets US AI policy through export controls, executive orders and agency action rather than standing statute.
European Union
The world's most comprehensive AI rule-maker, chiefly through the risk-based AI Act, the first statutory framework for artificial intelligence.
UK AI Security Institute
State-run evaluator that tests frontier AI models for serious security-linked risks; the first national AI safety institute of its kind.
Chinese Government
Regulates AI mainly through the Cyberspace Administration of China, weighted toward content control, provider registration and AI-content labelling.
UK Government
Pursues an adoption- and infrastructure-led AI strategy, leaning on state evaluation capacity and voluntary commitments rather than binding law.
State of California
US state whose AI laws function in practice as national regulation, given the concentration of major AI developers headquartered there.
US Center for AI Standards and Innovation
The US federal government's in-house evaluator of advanced AI, housed within NIST; focused on standards, innovation and the capabilities of foreign models.
G7
Coordinates AI governance among industrial democracies, notably through the Hiroshima AI Process and its voluntary code of conduct for advanced AI.
OECD
Author of the 2019 OECD AI Principles and host of the OECD.AI policy observatory; sets non-binding international standards for trustworthy AI.
United Nations
Convenes global AI norm-setting through non-binding resolutions and advisory panels rather than enforceable rules.
Academic labs & institutes
University research groups that generate the methods, benchmarks and people.
Stanford HAI
Stanford's human-centred AI institute, best known for the annual AI Index, a widely cited survey of the field's progress, investment and policy.
UC Berkeley
A leading academic source of AI methods, benchmarks and researchers, home to the Berkeley AI Research (BAIR) lab.
Tsinghua University
China's leading university for AI research, whose labs and alumni seeded much of the country's large-model industry.
MIT
Home to CSAIL, MIT's flagship computer-science and AI laboratory, and a long line of foundational work in AI and robotics.
Carnegie Mellon University
A US university with top-ranked programmes in machine learning, AI and robotics, including the world's first departments in robotics and machine learning.
Mila
Montreal deep-learning institute founded by Yoshua Bengio, one of the largest academic concentrations of AI researchers.
Alan Turing Institute
The UK's national institute for data science and AI, founded in 2015 as a broad university research hub and, since 2025, pushed by government toward a defence and national-security focus amid funding threats, restructuring and staff unrest.
George Williamson~£42m founding government grant (2015); a £100m five-year settlement (2024) that ministers have tied to a national-security refocus
ETH Zurich
Switzerland's leading science and technology university and a major European centre of AI and machine-learning research.
Vector Institute
Toronto deep-learning research institute, part of Canada's national AI strategy, with roots in Geoffrey Hinton's Toronto group.
Big tech & backers
Platform companies and financiers whose capital and clouds underwrite the frontier.
Microsoft
Its OpenAI partnership and its Copilot assistants across Azure, Windows and Office make it one of the central players in the current AI wave, though a 2026 deal made the OpenAI tie less exclusive.
Satya Nadella~$3.7tn market cap (2026); public (Nasdaq: MSFT)
Amazon
Sells AI infrastructure through AWS, is Anthropic's largest financial backer, and builds its own Trainium chips and the Bedrock model marketplace — positioning as a supplier of AI capacity rather than a frontier-model maker.
Andy Jassy~$3.0tn market cap (2026); public (Nasdaq: AMZN)
SoftBank
Acts as a financier rather than a builder of the AI boom — leading OpenAI's $40bn round, backing the Stargate project, and owning the chip designer Arm.
Masayoshi SonInvestment holding group; controls chip designer Arm; public (TSE: 9984)
Oracle
Repositioned its enterprise-cloud arm (OCI) as a major supplier of AI compute, signing a reported ~$300bn deal with OpenAI and taking a central role in the Stargate data-centre project.
Clay Magouyrk, Mike Sicilia, Larry Ellison~$420bn market cap (2026); public (NYSE: ORCL)
Salesforce
Enterprise software whose Agentforce platform builds autonomous AI agents for customer service and sales inside its customer-relationship-management products.
Marc Benioff~$160bn market cap (2026); public (NYSE: CRM)
IBM
Enterprise AI through the watsonx platform and the open Granite foundation models, plus a large consulting business helping firms deploy AI.
Arvind Krishna~$220bn market cap (2026); public (NYSE: IBM)
Media & rightsholders
Publishers and studios that appear chiefly as copyright litigants.
The New York Times
The New York Times, in this record chiefly as the news publisher that took OpenAI to court over training on its journalism and became the leading test of AI and copyright.
Universal Music Group
The world's largest music label; lead plaintiff against AI song generators Suno and Udio, and later Udio's partner in a licensed AI-music platform.
The Walt Disney Company
Entertainment conglomerate whose character library made it a lead studio plaintiff against Midjourney, and later a content partner to OpenAI's Sora.
Getty Images
Getty Images, a large stock-photography company that in this record features mainly as the rightsholder suing image generators, most prominently Stability AI in the US and UK.
News Corp
Murdoch news publisher (WSJ, New York Post, The Times) that licensed its journalism to OpenAI while its Dow Jones unit sued Perplexity over alleged copying.
Reddit
US discussion platform that both licenses its user content to AI developers and sues firms it says scraped it without permission.