The Databases That Keep the Score

Crunchbase, PitchBook, and CB Insights are the three primary sources behind nearly every round size, valuation, and unicorn count cited in startup journalism and academic research. None of them is complete, and none uses an identical method to build its records.

Crunchbase launched in 2007 as a directory embedded inside TechCrunch, tracking the companies and people the blog was already covering. AOL spun it out as a standalone business in 2015. Its data collection has always blended self-reporting — founders and investors submit their own rounds — with editorial curation and machine ingestion. That crowdsourced foundation makes Crunchbase the most permissive of the three: coverage is wide but uneven, skewing toward US markets and toward rounds large enough to generate a press release.

PitchBook was founded the same year, 2007, by John Gabbert in Seattle, and built around a research-analyst model rather than self-submission. Morningstar acquired it in 2016 for approximately $225 million, a figure disclosed in Morningstar's annual filings. PitchBook draws on SEC filings, regulatory disclosures, LP documents, and direct sourcing, which gives it stronger coverage of later-stage and private-equity transactions and more reliable historical data — at a subscription cost that puts it out of reach for most independent researchers.

CB Insights launched in 2008, founded by Anand Sanwal and Jonathan Sherry in New York. It runs a similar analyst-driven collection model but has built its differentiation around pattern recognition: exit timelines, investor scoring, and the sector-level aggregates that populate the quarterly venture reports widely cited by journalists. CB Insights was among the first to systematise unicorn tracking after Aileen Lee coined the term in November 2013.

All three share a structural limitation: private companies are not required to disclose valuations, and many rounds — particularly at the seed stage — are never reported publicly. The figures that circulate in coverage are therefore a floor, not a ceiling. Researchers at the National Venture Capital Association have noted the gap explicitly in methodology disclosures accompanying their annual data releases.