PitchBook becomes data partner for OpenAI's ChatGPT for Financial Services, delivering firmographic data via Essential connector
On 10 September 2026, PitchBook (a Morningstar unit) became a data partner for OpenAI's ChatGPT for Financial Services, with its Essential firmographic dataset on companies, investors and funds queryable inside the GPT-6 Astra product. Answers cite back to PitchBook, keeping the source visible to analysts even though the workspace and enterprise contract sit with OpenAI. Commercial terms were not disclosed.
Where the rake sits
PitchBook keeps subscription revenue from entitled seats and gets citation visibility inside every grounded answer; OpenAI keeps the workspace, the enterprise contract, and the renewal decision. The economic split between the two is not disclosed.
What happened
- Announced 10 September 2026 via Business Wire; PitchBook named a data partner for ChatGPT for Financial Services, OpenAI's tailored offering for investment banking and equity research teams built on GPT-6 Astra.
- The Essential connector exposes firmographic records on companies, investors and funds; the dataset is indexed within ChatGPT for Financial Services so no separate connector activation is required.
- Answers appear inside ChatGPT alongside citations back to PitchBook, letting teams inspect supporting evidence while building models, pitchbooks and client materials.
- PitchBook is a Morningstar business unit (NASDAQ: MORN); Morningstar acquired PitchBook in 2016 (reported enterprise value ~$225M) and PitchBook now serves more than 100,000 clients from 3,000+ employees across Seattle, San Francisco, New York, London, Singapore and Mumbai.
- PitchBook runs parallel partnerships with Anthropic, Google, Microsoft, Perplexity and Rogo, and Morningstar announced a Gemini Enterprise for Financial Services integration in August 2026 — this is one of several, not an exclusive.
- Commercial terms (fee, revenue share, exclusivity, term) were not disclosed in the release.
Who is involved
Seattle-based private-capital-markets data platform (a Morningstar business unit); the data asset here is its Essential firmographic dataset on companies, investors and funds, exposed inside ChatGPT with citations.
Founded 2007; part of Morningstar since 2016 (acquired at roughly $225M enterprise value); serves more than 100,000 clients worldwide per its own release.
Independent investment research group; parent of PitchBook, which is one of its five reportable segments alongside Data & Analytics, Wealth, Credit and Retirement.
NASDAQ: MORN; market cap approximately $7.17B and trailing-twelve-month revenue $2.57B as of September 2026; roughly 10,973 employees.
Operator of ChatGPT for Financial Services, the buyer-facing product into which PitchBook's Essential dataset is delivered; embeds partner data behind a GPT-6 Astra reasoning experience for finance teams.
Private company; valuation and revenue figures widely reported in the press but not audited.
The reading
The enrichment sits on both sides but is split cleanly: PitchBook maintains the firmographic substrate on companies, investors, funds, deals and people, while OpenAI's ChatGPT for Financial Services combines that feed with GPT-6 Astra reasoning to produce research, financial models and client materials. From PitchBook's seat the work it does is on the substrate itself, not on the end-user analyst's decision.
The payer here is the ChatGPT for Financial Services subscriber - a finance team deciding, for example, which private company to diligence or which fund to approach next - and PitchBook's value to them is that its coverage of private companies, investors and funds is the one input GPT-6 Astra cannot generate from public text alone. The substrate's enrichability rests on being the authoritative private-markets reference the model is grounded against.
What crossed is queryable access to PitchBook's expanded firmographic dataset through the Essential connector inside ChatGPT; what did not cross, on the public facts, is the underlying PitchBook Platform relationship or any transfer of the raw database - the dataset is reached in place via a connector. This is the Tempus/Gilead pattern: delivery through someone else's surface does not make the dataset the surface's property.
Not determinable. Financial terms were not disclosed, and with OpenAI holding the end user and the reasoning layer, PitchBook's exposure to under-capture depends on connector economics that are not public.
Why it matters
One angle worth testing: this is the same shape as the earlier PitchBook/Quartr-into-ChatGPT ruling — the dataset is the deliverable even though the buyer never leaves OpenAI's chat surface, and the citation requirement is what keeps PitchBook visible to the analyst rather than fading into GPT-6 Astra's answer. A second angle: the parallel deals with Anthropic, Google, Microsoft, Perplexity and Rogo suggest PitchBook is deliberately refusing to let any one LLM buyer become the line to the analyst — worth asking whether that discipline is what protects the rake here.
The argument this deal tests: The rake is decided before the negotiation begins
Related deals
- Customers Bank signs OpenAI deal to automate lending, deposits and payments using its own operational data2026-04-27
- Circana launches Liquid Mix with Google Meridian, combining Google's MMM framework with Circana's first-party data2026-09-08
- Owkin licenses K Pro AI scientist and MOSAIC multimodal patient data to Servier for oncology research2026-09-11
- Tempus AI expands multi-year collaboration giving Gilead enterprise-wide access to Lens platform and multimodal real-world oncology data2026-04-09
Sources
- finance.yahoo.com — primary
- pitchbook.com — party background
- stocktitan.net — party background
- pitchbook.com — party background
- uk.finance.yahoo.com — party background
Announced 2026-09-10