AI Companies
AI Companies is a data page on the AI desk of Glassbox Auditing. The companies behind the models, public and private on one footing: status, models tracked, capability, compute, SEC Form D funding and reported valuations. Private labs such as OpenAI and Anthropic set the pace of the field while disclosing far less than listed companies.
Last updated: .
What is on this page today?
- Amazon is a documented compute or cloud supplier to 8 AI developers, Oracle to 4 — 11 providers supply two or more developers · links read from the companies' filings and Epoch's supercomputer record · a supplier also sells to others; documented, not exclusive (SEC filings and Epoch AI Supercomputers, 7 Nov 2026)
- 49 companies develop the models this desk tracks; Alibaba has the most (126) — 21 private, 12 public, 3 subsidiary · models tracked, not a ranking on figures some companies do not publish · status from the public record (listed, owned by a parent, nonprofit) (Wikidata and Epoch AI, 7 Nov 2026)
- OpenAI has the most documented investors and owners among AI developers: 6 organisations — Microsoft, SoftBank Group, Sequoia Capital, Andreessen Horowitz and 2 more · 18 developers have at least one · documented stakes from Wikidata and SEC filings, not claims of influence or control (Wikidata and SEC filings, 6 Nov 2026)
- BlackRock, Inc. holds a documented stake in 4 AI developers, Vanguard Group in 4 — 5 organisations hold stakes in two or more developers · 2 people are directors or executives at two or more · documented links, not claims of influence or control (Wikidata and SEC filings, 6 Nov 2026)
- 3 listed companies hold a documented stake in Anthropic, the private AI developer with the most — Amazon, Salesforce, Alphabet Inc. · stakes read from Wikidata and filings; a missing share is not reported (Wikidata and SEC filings, 3 Nov 2026)
- xAI: 592,180 H100-equivalents of AI supercomputers on record, 575,796 added in the year to Jul 2025 — 4 clusters it owns or trains on (a shared one counts for each company) · Epoch's record runs to Jul 2025 (Epoch AI Supercomputers, 21 Jul 2025)
- Meta: 161,738 H100-equivalents of AI supercomputers on record, 100,000 added in the year to Jul 2025 — 9 clusters it owns or trains on (a shared one counts for each company) · Epoch's record runs to Jul 2025 (Epoch AI Supercomputers, 21 Jul 2025) META
- OpenAI: 108,514 H100-equivalents of AI supercomputers on record, 100,000 added in the year to Jul 2025 — 3 clusters it owns or trains on (a shared one counts for each company) · Epoch's record runs to Jul 2025 (Epoch AI Supercomputers, 21 Jul 2025)
- Microsoft: 132,934 H100-equivalents of AI supercomputers on record, 100,000 added in the year to Jul 2025 — 18 clusters it owns or trains on (a shared one counts for each company) · Epoch's record runs to Jul 2025 (Epoch AI Supercomputers, 21 Jul 2025) MSFT
- Nvidia: 24,029 H100-equivalents of AI supercomputers on record, 2,048 added in the year to Jul 2025 — 17 clusters it owns or trains on (a shared one counts for each company) · Epoch's record runs to Jul 2025 (Epoch AI Supercomputers, 21 Jul 2025) NVDA
- Google: 14,763 H100-equivalents of AI supercomputers on record, 0 added in the year to Jul 2025 — 9 clusters it owns or trains on (a shared one counts for each company) · Epoch's record runs to Jul 2025 (Epoch AI Supercomputers, 21 Jul 2025) GOOGL
Where does the data come from?
Items on this page come from SEC filings and Epoch AI Supercomputers; Wikidata and Epoch AI; Wikidata and SEC filings; Epoch AI Supercomputers. Each item links to its own source. All sources, how items are checked.
What does this page lead to?
- AI Supply Chain (AI): Money raised by AI developers is mostly spent on compute: chips, cloud contracts and data centers, which is revenue for the supply chain's layers. (moves the same way; quarters; sector scale) Source: SEC Form D data sets: private offerings under Regulation D (DERA, 2008 on).
- Job Postings (Feed): Developers put part of new funding into hiring; their open roles appear in the AI jobs feed, read from their own career pages. (moves the same way; months; company scale) Source: artificialintelligencejobs.co jobs API (live jobs at AI companies, from their career pages).
- Regulatory Exposure (Companies): The bigger a developer's models and user base, the more AI laws apply to it (the EU AI Act's general-purpose model duties, state rules) and the more regulators' matters resemble its situation. (moves the same way; months to years; company scale) Source: Regulators' own announcements: FTC, SEC, Justice Department, EEOC, state attorneys general, European data-protection authorities (each matter links its own).
- Cloud & Hubs (Companies): The big labs are anchor customers of the clouds (OpenAI with Microsoft and Oracle, Anthropic with Amazon and Google), so their growth adds to cloud usage and contracts. (moves the same way; quarters; company scale) Source: AWS and Google Cloud customer stories.
- Nation (AI): Developers' and their cloud partners' purchases of chips, servers and software count as business investment in GDP, the tech investment the Nation page tracks. (moves the same way; quarters; economy scale) Source: U.S. Bureau of Economic Analysis (NIPA).
- AI Researchers (AI): New funding pays for hiring; when a developer takes on a tracked researcher, the move shows in that author's OpenAlex affiliation on their next papers. (moves the same way; months; company scale) Source: OpenAlex (scholarly works).
- Technology (Global Intelligence): Money raised by AI developers pays for new training clusters, which appear on the Technology map as AI data centers from Epoch AI's supercomputer list, each with its owner. (moves the same way; quarters; sector scale) Source: Epoch AI, AI Supercomputers (CC BY 4.0).
What feeds this page?
- Labs & Countries (AI): Each lab's best score is the capability column on its developer's profile, beside its funding and compute. (moves the same way; same day; company scale) Source: Epoch AI Capabilities Index (CC BY 4.0).
- AI Law & Enforcement (AI): Duties such as the EU AI Act's model documentation, risk assessment and incident reporting add compliance costs for model developers, as set in each law's own text. (moves the same way; years; company scale) Source: Official texts: EUR-Lex, state legislatures, the Federal Register and national gazettes (each law links its own).
- Job Listings (Companies): Open roles at AI companies are where new funding goes first, so a lab's postings tend to rise after it raises money. (moves the same way; months; sector scale) Source: artificialintelligencejobs.co jobs API (live jobs at AI companies, from their career pages).
- Cloud Outages (Global Intelligence): Outages at AI model providers' APIs stop the apps and agents built on them for as long as they last, a dependency that grows as more companies route work through a few labs. (either way, depending on the cause; same day; sector scale) Source: OutageDeck public API (cloud and SaaS provider status, normalized from official status pages).
- H-1B & Talent (Immigration): AI labs and the big tech firms that fund them sponsor many foreign-born researchers and engineers, so H-1B rules and fees bear on their hiring costs. (either way, depending on the cause; months; company scale) Source: USCIS H-1B Employer Data Hub.