Two states legalised psilocybin by ballot. The money came from one side in both.

A committee total measures paperwork on one side and says nothing about how hard the other side tried. With that in front:

Colorado’s campaign-finance register reports $5,807,698.30 of contributions across the two committees supporting the 2022 natural-medicine measure, against $50,986.00 for the committee opposing it. Massachusetts reports $7,752,198.94 of receipts for the committee supporting Question 4 in 2024, against $137,072.50 opposing.

Those two pairs are not one number. Colorado reports contributions and Massachusetts reports receipts, gathered over different windows, and no proportion between any two of these figures appears on this page.

Of the supporting committee’s own $5,789,588.30 in Colorado, which is a single committee and not the two-committee total above, $5,700,187.15 came from organizations against $89,401.15 from individuals, with $110.00 attributed to no class at all. Massachusetts itemises a wider window: of the $8,998,923.26 its register attributes to a contributor class, $4,746,492.26 came from organizations and $4,252,431.00 from individuals. That itemised total is not the $7,752,198.94 above counted twice; it is a different measure.

So one campaign was funded almost wholly by organizations and the other roughly half by people. The word organizational does not carry across that pair. And two states is a sample of two: Colorado’s shape will be read as typical because it is the more striking, and nothing in either register makes it typical of anything.

Those four spellings are one contributor, written four ways by whoever was filling in the form.

A contributor string is a name and not an identity, and until somebody reconciles the spellings by hand the register cannot answer the question people actually ask it, which is who gave what. The totals above are sums and are solid. The attribution underneath them is as good as the typing.

This piece does not name that contributor, or any other. Naming the largest organizational giver in a state from a register that cannot reliably distinguish one giver from another would be asserting an identity the source does not establish.

Colorado’s lobbying register is the only one of the five this project holds that answers a query cleanly. On the 7 September 2026 read it carried 530 registrations naming a natural-medicine or psilocybin subject across four fiscal years, which fold down to 49 client organizations once the names are reconciled. Two hundred and ninety of those rows carry the declared position Monitoring, which is a lobbyist telling the state they are watching. Eight of the 49 clients attach to the value chain at all. The rest are hospitals, counties, cities, professional bodies and sheriffs.

Registration is not influence, and the register says so itself in the position field. Run the same query on the word psychedelic and it returns nothing, which is a fact about the vocabulary the register uses and not about the field.

And this is one state. Oregon’s Measure 109 money and the federal lobbying half are both unread for this work, so nothing here describes a national picture of policy money.