Wiki · AI Players · Index · Last reviewed June 19, 2026

AI Organizations

A neutral index for institutions shaping the AI ecosystem: frontier labs, cloud and chip firms, data and labor suppliers, safety institutes, standards bodies, civil-society organizations, public agencies, and application companies. The page is a map of power and accountability, not a ranking or endorsement.

Category: Index Published: June 19, 2026 Modified: June 19, 2026 Last reviewed: June 19, 2026 Tags: Organizations, Governance, Infrastructure, Standards, Frontier AI

Definition

An AI organization is an institution that materially affects how AI systems are built, deployed, evaluated, financed, governed, or contested. The category includes companies, university institutes, nonprofits, standards bodies, public agencies, industry consortia, civil-society groups, and infrastructure firms.

The useful question is not whether an organization "does AI" in a marketing sense. The useful question is what role it plays in the AI stack: model development, compute, data, tooling, product distribution, evaluation, standards, regulation, labor, public accountability, or rights protection.

For governance, an organization is not only a brand. It is a legal entity, funding structure, decision process, infrastructure dependency, jurisdictional footprint, and accountability surface. A lab, cloud provider, standards body, regulator, school vendor, or data-labeling firm can all shape the same deployed AI system.

Scope

This page is an index, not a ranking. It groups organizations by their role in the AI ecosystem and points to individual profiles where the site has them. Organizations should be added conservatively, with dated sources and clear relevance to AI capability, governance, infrastructure, labor, safety, civil rights, public services, or public-interest technology.

The index should not become a directory of every startup using AI language. Inclusion is strongest when the organization has durable influence over frontier capability, widely used models or tools, AI infrastructure, public standards, legal obligations, evaluation practice, labor supply chains, or social systems where people need accountability and recourse.

Entries should preserve boundaries between an organization's own claims, independent evidence, and public obligations. A press release can establish that an organization announced a partnership, model, safety framework, or funding round. It does not by itself prove real-world safety, market durability, regulatory compliance, or social benefit.

Current Context

As of the June 19, 2026 review, AI organizations sit inside overlapping governance regimes rather than one unified AI law. The EU AI Act is in staged rollout: general-purpose AI rules began applying on August 2, 2025, and the European Commission timeline lists August 2, 2026 for most rules, Article 50 transparency duties, innovation measures, and enforcement to start applying.

In the United States, federal AI policy is organized around acceleration, infrastructure, standards, procurement, and security. The official AI Action Plan site describes three pillars: accelerating innovation, building AI infrastructure, and leading in international diplomacy and security. NIST's Center for AI Standards and Innovation, or CAISI, describes itself as the U.S. government's industry contact point for testing and collaborative research on commercial AI systems.

International references also shape what organizations are expected to document. The OECD AI Principles were adopted in 2019 and updated in 2024, while the Council of Europe Framework Convention on Artificial Intelligence opened for signature in 2024 as a legally binding treaty focused on human rights, democracy, and the rule of law.

Standards and evaluation institutions are becoming more important. NIST's AI Risk Management Framework remains a central voluntary risk-management reference; ISO/IEC 42001 supplies a management-system standard for organizational AI governance; NIST's 2026 AI Agent Standards Initiative treats agent identity, interoperability, and secure action as standards problems. Industry coordination also matters: the Frontier Model Forum lists Amazon, Anthropic, Google, Meta, Microsoft, and OpenAI as current members.

The 2026 International AI Safety Report notes that public information about how leading general-purpose AI systems are built and evaluated is often scarce, while most risk-management initiatives remain voluntary. Stanford HAI's 2026 AI Index says industry produced more than 90% of notable frontier models in 2025 and that responsible-AI measurement is not keeping pace with capability. That makes organizational accountability central: the strongest evidence often sits inside the same institutions that have incentives to ship.

The organizational map is also widening. Model developers remain central, but infrastructure firms, cloud providers, chip suppliers, data centers, evaluation bodies, app distributors, schools, hospitals, courts, public agencies, and labor vendors increasingly determine how AI reaches people. A serious AI organization entry must therefore track role, dependencies, and deployment context, not only model names.

Index Groups

Profiles

Governance, Standards, and Evaluation

Model Developers and Platforms

Infrastructure, Data, and Application Layers

Governance Implications

AI governance has to follow organizations, not just models. A deployed system may depend on a frontier lab, cloud provider, chip supplier, data broker, annotation vendor, product platform, enterprise integrator, standards body, and public agency at once. Accountability weakens when each actor points to another part of the stack.

Responsibility mapping: organization profiles should identify who trains the model, hosts inference, supplies compute, curates data, operates the user interface, sells the product, audits the system, and controls updates. Without that map, incident response becomes a chain of denials.

Concentration: frontier capability, compute, cloud distribution, app defaults, and safety evidence are increasingly concentrated in a small number of organizations. That concentration can improve coordination, but it also raises competition, capture, and public-dependence risks.

Voluntary governance: industry forums, model cards, safety frameworks, and private evaluations can produce useful evidence. They do not substitute for public authority, independent access, enforcement, or recourse for affected people.

Infrastructure as control point: chips, cloud accounts, model weights, data centers, identity systems, app stores, payment rails, and enterprise procurement can become practical levers of AI governance. Controls at these layers can also entrench incumbents if they are designed without access, auditability, and competition safeguards.

Application-layer harm: some of the most immediate governance questions arise when AI is embedded into schools, workplaces, health systems, public benefits, policing, finance, search, and companionship. Organization profiles should therefore track deployment setting and recourse, not only model capability.

Public-private boundary: public agencies increasingly buy, test, regulate, subsidize, or rely on private AI systems. Profiles should note when an organization is a public authority, a vendor to public authorities, a standards participant, or a private actor setting de facto rules through market power.

Entry Standard

Organization entries should identify what the organization does, why it matters, where it sits in the AI stack, which jurisdictions or public obligations matter, what evidence supports its claims, and which risks or governance questions surround it.

For fast-changing companies, entries should avoid treating strategy, valuation, product availability, partnerships, membership lists, leadership structure, model access, compute supply, customer counts, or safety commitments as timeless facts. Date those claims and prefer primary sources: official announcements, regulator publications, standards documents, court filings, technical reports, annual reports, system cards, or published research.

Entries should also distinguish marketing language from observed behavior. Do not add unsupported hype, AGI claims, divinity claims, consciousness claims, or prophecy. If an organization makes such claims or if communities make them around a product, the page should describe that as a belief, narrative, or controversy with sources, not as a fact about the system.

Source Discipline

Organization claims should be sourced according to the kind of claim. Use corporate pages and announcements for self-descriptions, product launches, membership, funding, partnerships, and governance frameworks. Use regulator publications, court filings, annual reports, official standards, public procurement records, and technical papers for obligations, ownership, legal duties, audits, and performance claims.

Do not treat a company safety framework, benchmark table, model card, or policy post as independent proof that the organization is safe or accountable. It is evidence of what the organization says it tested, valued, or committed to at a specific date. Stronger entries explain who could verify the claim, what evidence is missing, and what would change the assessment.

Membership lists, leadership pages, product access, model availability, valuations, strategic partners, and public-agency names can change quickly. Each should carry a review date or source date. When an organization is included because of controversy, enforcement, labor conditions, civil-rights impact, or public-sector deployment, use primary records where possible rather than secondhand summaries.

Sources

Return to Wiki