# OpenAI and Anthropic in public company disclosures

Inspect six provider observations across three companies, including co-disclosure and the limits of a small public-source sample.

Six observations, three companies. Historical snapshot reviewed September 10, 2026; source pages rechecked September 12. No market-share or spending claim.

## The observed sample

The September 10, 2026 reviewed snapshot contains Intercom, Linear, and Sentry disclosures naming both OpenAI and Anthropic. That is six company-provider observations and three unique companies. It is a deliberately small coverage sample, not a representative estimate of market share.

Company	OpenAI named	Anthropic named	Observation date
Intercom	Yes	Yes	2026-09-10
Linear	Yes	Yes	2026-09-10
Sentry	Yes	Yes	2026-09-10

## Count companies and relationships separately

Each company can appear under both providers. The provider columns therefore overlap and should not be added to claim six customers. In this sample, all three companies co-disclose both providers. That describes the chosen sample only.

These records name the provider for a disclosed purpose. They do not establish equal usage, direct billing, contract size, model preference, or company-wide deployment. Read the disclosure’s current region and scope before applying the observation to a particular customer.

## What the comparison does not show

Co-disclosure does not mean one provider replaced the other. The sample has no representative historical panel, so this page makes no claim about growth, churn, market share, or competitive wins.

Public disclosures can lag operational changes and can include optional processing arrangements. Absence from this sample means only that the company is not among these reviewed records.

## Reproduce and extend the sample

Download the six records with source URLs, scope, delivery labels, and observation dates. Use domain plus provider as the relationship key and domain alone for unique-company counts. Preserve new observations with their own dates.

To research movement, first collect comparable earlier evidence for the same company and scope. Do not convert the date you first found a page into an adoption date. A larger sample needs a defined inclusion method before it can support broader conclusions.

## Checklist

[ ] Use unique companies as the denominator.
[ ] Allow provider columns to overlap.
[ ] Read scope and delivery details.
[ ] Keep co-disclosure distinct from replacement.
[ ] Do not extrapolate this sample to market share.

Notes:


Intercom subprocessor disclosure: https://www.intercom.com/legal/subprocessors-list
Linear data-processing agreement: https://linear.app/dpa
Sentry subprocessor disclosure: https://sentry.io/legal/subprocessors/