How the ranking is made.

Who is listed. A company is here when its own website says it sells datasets to the companies that build AI: 115 companies, each read on its own pages, with the sentence that says so quoted and linked.

How they are ranked. Each company is ranked on the first of these it has on record: its revenue over twelve months (a run rate or a year as published, or its last four quarters added up; an AI segment where the rest of its business is not AI data), then its newest valuation, then the money it disclosed raising. No company pays for a place.

Where the figures come from. Every revenue figure, valuation and round links to the filing or article that stated it. A company with no figure on record is listed by name, never ranked by guess.

Who publishes it. AI Training Data Providers is published by the team behind AIData.Broker, which keeps the records it is built from. The same records, ranked the same way, are on AIData.Broker's own ranking of AI training data providers, and every table is free to reuse as open data that we publish there. No provider pays to be listed or ranked.

Records last reviewed 8 Nov 2026.