The Open Foundation List, data-2026-10-08

The Open Foundation List, first release (data-2026-10-08)

Every U.S. private foundation that files Form 990-PF electronically, as small CSV files you can open in a spreadsheet or load into a database.

Built from IRS e-file index years 2017 to 2026 (fiscal years 2014 to 2026), the IRS Exempt Organizations master file, the IRS Automatic Revocation list (dated 2026-09-30) and Publication 78 (dated 2026-09-10). Every file was read from the database in one snapshot on 2026-10-09 at 08:32 UTC.

File	Rows	What it is
foundations.csv.gz	166,036	One row per foundation: name, place, latest numbers, what its latest return says about applications, IRS standing.
foundation_years.csv.gz	914,010	One row per foundation and fiscal year.
foundation_grants_<year>.csv.gz (2014 to 2026)	5,603,910 in total	One row per grant whose recipient is linked to an organization record. Other grants are counted in foundation_years.csv.gz, not named.
manifest.json		Row counts, sha256 of every file, sources, the checks that ran and their results.
README.md		The column dictionary and the reading rules.
LICENSE.txt		CC BY 4.0 and the credit line.

Application posture (from the latest return): 27,880 say they accept applications, 105,099 say they give only to preselected organizations, 33,057 do not state it.

Checks. Boundary checks F1 to F8 and pre-flight checks P1 (no amended return counted twice) and P2 (application answers as new as the financial data) all passed. Results are in manifest.json.

Left out on purpose. The purpose text of a grant (it can name a private person), street addresses, and any contact that is not a public role-based channel. A missing value is empty, never 0. "Not stated" is not "closed".

Licence. The compilation is CC BY 4.0. Credit: "Open Funder Database contributors (GetFunded), from IRS e-file data". The underlying records are public records of the U.S. government.

Browse the same data at https://getfunded.ai/foundations. Rebuild it yourself with funderdb export foundations in corpus/.