Commercial developers build nearly every notable AI model, and the money behind them is concentrated in the United States. The Stanford AI Index 2026 puts corporate AI investment at a record USD 581.69 billion in 2025. No official statistical category yet exists for "AI companies", so every count of them depends on how a researcher defines the term.
Key takeaways
- Newly funded AI companies rose 71 percent in 2025, and billion-dollar funding events nearly doubled.
- Corporate AI investment was USD 581.69 billion in 2025, split into USD 344.66 billion of private investment and USD 214.44 billion of mergers and acquisitions.
- Frontier labs have stopped reporting some technical details of their largest systems, so outside counts of what they build are partly estimates.
- A July 13, 2026 federal notice on the 2027 revision of the North American Industry Classification System (NAICS) mentions artificial intelligence once, as a topic a single commenter raised, and recommends no industry with that name.
Who builds the most capable models
The AI Index is an annual report from the Stanford Institute for Human-Centered AI. Its 2026 research and development chapter credits companies with more than nine in ten of the notable models released in 2025. Universities and joint projects make up the remainder, though the chapter gives no separate academic count. Country comparisons are in the same chapter, which also says China leads in publication volume.
The chapter also describes a change in how much companies publish about their own work. Reported parameter counts have stayed near one trillion for three years, yet the report notes that reporting from frontier labs has stopped, while independently estimated training compute has kept rising. A count of "companies that build large models" is therefore a count of organizations whose output can be observed, not a full census.
Where the money went in 2025
The economy chapter of the 2026 report tracks funding rather than revenue. Corporate AI investment grew from USD 14.57 billion in 2013 to USD 581.69 billion in 2025. Private investment grew 127.5 percent in the latest year and accounts for 60 percent of the total, with the rest coming from mergers and acquisitions. Newly funded AI companies rose 71 percent, billion-dollar funding events nearly doubled, and generative AI took nearly half of all private AI funding.
| Measure (2025) | Amount | Note |
|---|---|---|
| Total corporate AI investment | USD 581.69 billion | Private investment plus mergers and acquisitions |
| Private investment | USD 344.66 billion | Up 127.5 percent, 60 percent of the total |
| Mergers and acquisitions | USD 214.44 billion | Remaining share of the total |
| Private investment, United States | USD 285.88 billion | About 23 times the Chinese figure |
| Private investment, Europe | USD 20.92 billion | Regional figure as shown in the chapter |
| Private investment, China | USD 12.41 billion | Likely understates total Chinese funding |
The chapter itself adds the caveat on China: government guidance funds are estimated to have put USD 184 billion into AI firms between 2000 and 2023, and that money does not appear in a private-investment total. Canada is not broken out in the chapter text reviewed for this page.
What these counts leave out
Investment totals measure what investors paid, not what companies earn or ship. A firm that raises a large round and a firm that profitably sells a narrow product both count as one company, and a large acquisition of a small team can weigh more than years of private funding. Model counts have a similar limit: they cover systems that meet the report's definition of notable, so a company that applies existing models inside a hospital or a logistics network may never appear in a model tally.
Adoption figures describe a different group altogether. The same chapter reports that organizational adoption of AI rose from 55 percent in 2023 to 88 percent in 2025, and that 70 percent of organizations use generative AI in at least one business function. Those percentages count organizations that use AI, which is a much larger set than the organizations that build it.
Why no official list of AI companies exists
Government statistics group businesses by what they produce. The Office of Management and Budget notice published in the Federal Register on July 13, 2026 describes NAICS as grouping establishments that use the same or similar production processes. The notice proposes updates for 2027, to apply to establishment data for periods beginning on or after January 1, 2027, and it follows a December 20, 2024 request that drew 60 submissions.
Among those submissions, the notice says, single comments asked for various new industries or clarifications, and the list of topics includes artificial intelligence. The recommended changes that follow include new industry content for media streaming and for social networks and collaborative media sites. The notice names no industry for artificial intelligence, and its only mention of the term is that list of comment topics. Until a code exists, researchers build their own definitions, which is why two reports can count different numbers of AI companies in the same year.
Checklist for reading any count of AI companies
- Which definition is used: builders of models, sellers of AI products, or any firm that mentions AI?
- Is the figure a count of firms, a funding total, or a model tally?
- Does the source say whether acquisitions, government funds, or non-public firms are included?
- Which year and which data cutoff does it cover? The AI Index page states data as of March 2026.
- Is the source an independent report or the company's own marketing?
FAQ
How many AI companies are there?
No official figure exists. The sources reviewed for this page report notable models and funding amounts, and the AI Index says newly funded AI companies rose 71 percent in 2025 without giving a total count in the chapter text reviewed.
Are most AI models built by companies or universities?
By the AI Index 2026 count, companies. Industry produced over 90 percent of notable AI models in 2025.
Does investment size show which companies are best?
No. Investment totals show how much capital was committed, and the report does not rank companies by quality or usefulness.
Why do different reports give different numbers?
Reports differ in definition, data source, and cutoff date. A figure is best cited with its edition and year.
Related reading
- How AI is used in data science
- AI use across technology, health care, and finance
- The Deloitte AI Institute explained
- Common AI technologies in use today
More in AI Tools and Industry.
Last verified: October 2026
Sources
- AI Index Report 2026, Research and Development chapter, Stanford Institute for Human-Centered AI, 2026
- AI Index Report 2026, Economy chapter, Stanford Institute for Human-Centered AI, 2026
- Statistical Policy Directive No. 8: North American Industry Classification System, Request for Comments on Proposed Updates for 2027, Office of Management and Budget, Federal Register, July 13, 2026