OpenAI accounts for 48.7% of documented links between UK companies and AI providers in a GreenPT study, according to AI Insider. The finding shows which providers appear most often in the public records collected, not how much AI businesses actually use.
OpenAI Watch analysis
What happened
GreenPT counted 1,377 company-provider links in public material such as company statements, supplier case studies and privacy notices. OpenAI appeared in 670, followed by Anthropic with 322, open-source models as a combined category with 235, and Google with 150.
These are links, not a count of separate businesses: a company naming two providers contributes two entries. Nor is the collection a representative survey of UK firms. Businesses that have not publicly identified their providers are missing from it.
Why it matters
The ranking gives a useful view of what turns up in public documentation, but it cannot tell us provider market share, spending or the scale of actual AI workloads. GreenPT also estimated that about 31% of the links involving OpenAI, Anthropic and Google lacked evidence meeting its standard for corroborated use. Some referred to proposals, pilots, outdated claims or conditional plans.
GreenPT assessed a separate sample of 734 links across 15 providers. It estimated that 1,023 of the full 1,377 links met its evidence standard, but that overall figure was projected from the smaller sample, not checked entry by entry.
Our read
OpenAI’s lead in public mentions is clear. What those mentions mean is less tidy: the study is a map of disclosures, not a census of adoption. That distinction is doing rather a lot of work here, as distinctions tend to when a ranking looks like a league table.
For readers assessing AI adoption, the useful question is not only which provider is named most often, but whether the document describes a live, evidenced deployment or merely an intention to use one.
What to watch
- Whether GreenPT publishes further detail on its collected records and evidence classifications.
- How the ranking changes when additional providers are included.
- Whether future research measures actual deployments, spending or workloads rather than public references.
Discussion spark: Should public mentions of an AI provider count as evidence of adoption, or should rankings include only confirmed deployments?
Sources and evidence
- OpenAI Most Often Named in U.K. Companies’ AI Disclosures, Study Finds – AI Insider (6 November 2026, 11:25 UTC)
OpenAI Watch is independently operated by WittyWires. It is not affiliated with, endorsed by, or operated by OpenAI.