Cannabis Industry Statistics — Live

Counts computed from our registry and cultivar data at page load, each with a stable anchor and a ready citation. The counting is continuous; the underlying records are not — each one carries the date we last pulled it, and those dates are published below.

How current is this, exactly?

We run a scheduled re-pull job. That is not the same as every record being checked every day, so we publish the record ages rather than the schedule.

  • Records pulled from source in the last 24h: 100% · 7 days: 100% · 30 days: 100%.
  • Most recent record pull: 2026-07-27 17:39:49. Oldest record still carried: 2026-07-27 17:37:31.
  • Last successful import run: 2026-07-27 17:39:20 (23 hours ago). Re-check target: daily (scheduled job).
  • The most recent import attempt failed. Affected sources keep their previous data and their ages keep climbing until a run succeeds.
  • Publisher update cadence: not established — registries publish on their own timetables and state no SLA we capture, so we do not infer one.
  • Source publication date per record: not established — these sources do not hand us a snapshot or effective date on ingest. What a record does carry is the registry's own issue and expiry dates, and the time we pulled it.
  • These counts were computed: 2026-07-28 17:06:12 (cached up to one hour). That is when we counted, not when the data was checked.
8,524

Cannabis license records tracked

Government-registry records in our database across all covered jurisdictions. One record is one license, not one business: a single company can hold many licenses across types and states, so this is always larger than a company count. Counted 2026-07-28 (see how current the underlying records are). Explore the underlying data →

Cite this statistic

🧩 Embed this as a live card →

21

US jurisdictions covered

States and districts with at least one record in our database. This measures how many sources we have wired, not how many markets exist — the number rises when we add a source, not when a market opens. Counted 2026-07-28 (see how current the underlying records are). Explore the underlying data →

Cite this statistic

🧩 Embed this as a live card →

94%

Share of records recorded as active

8,015 records carried an active-class status the last time each was pulled from its source; records pulled at different times are counted together, so this is a blended reading rather than a single-day snapshot. Denominator is all tracked registry records — marijuana, hemp, and every license class combined. Registry status is a public record, not a legal opinion on any license. Counted 2026-07-28 (see how current the underlying records are). Explore the underlying data →

Cite this statistic

🧩 Embed this as a live card →

21,375

Registry changes detected in the last 30 days

New licenses, status flips, and renewal signals our diff engine caught between pulls. It counts changes we detected, not every change that occurred: a change is only visible once the affected source is pulled again, and history is not backfilled before tracking began. Counted 2026-07-28 (see how current the underlying records are). Explore the underlying data →

Cite this statistic

🧩 Embed this as a live card →

5,256

Multi-license operator groups

Distinct operators holding more than one license, mapped across state lines. Counted 2026-07-28 (see how current the underlying records are). Explore the underlying data →

Cite this statistic

🧩 Embed this as a live card →

1,625

Cities with licensed activity on record

Unique city-and-state pairs appearing on any tracked registry record. Counts every status, not active-only, and depends on the city field being populated by the source. Counted 2026-07-28 (see how current the underlying records are). Explore the underlying data →

Cite this statistic
NY (2,936), TX (761), WA (514)

Largest state registries

The three biggest jurisdictions in our database by record count. Counted 2026-07-28 (see how current the underlying records are). Explore the underlying data →

Cite this statistic
9,222

Strain reference entries

Educational cultivar profiles with lineage, flavor, and FAQ data. No medical claims. Counted 2026-07-28 (see how current the underlying records are). Explore the underlying data →

Cite this statistic

🧩 Embed this as a live card →

234

Open industry roles listed

Live openings on the jobs board across all categories. Counted 2026-07-28 (see how current the underlying records are). Explore the underlying data →

Cite this statistic
19

Industry events published

Organizer-submitted or owner-verified B2B event entries published on the calendar. This is the count of published entries, not a count of future-dated events only. Counted 2026-07-28 (see how current the underlying records are). Explore the underlying data →

Cite this statistic
10,111

Mapped parent–offspring genetic links

Breeding relationships extracted from cultivar lineage records and resolved into a navigable genetics graph; 9,369 resolve to a catalogued parent cultivar. Counted 2026-07-28 (see how current the underlying records are). Explore the underlying data →

Cite this statistic

🧩 Embed this as a live card →

OG Kush

Most-crossed cannabis genetic

The single most frequently used parent cultivar in our catalogue — 248 catalogued crosses list it as a parent, across 1,540 cultivars that appear as a parent at least once. Counted 2026-07-28 (see how current the underlying records are). Explore the underlying data →

Cite this statistic

🧩 Embed this as a live card →

Methodology and what these numbers are not

  • A license record is not a business. One company can hold many licenses across types and states, and one premises can appear under more than one record. Record counts therefore run well above company counts and must not be reported as a number of businesses, operators, or dispensaries.
  • Coverage growth is source growth. Jurisdiction and record counts rise when we wire an additional state source or a source republishes more rows. That is our coverage expanding, not the market expanding. Absence of a state means absence of coverage, not absence of a market.
  • Counting is live; the records are not. Counts are computed from our normalized state-registry database at page load and cached for up to an hour. Each underlying record was last pulled from its source at its own time — see how current is this, exactly for the distribution of those ages. A scheduled import job is not evidence that any particular record was checked today.
  • Status is a copy, not a ruling. Every record shows its source label and the time we last pulled it. A status reproduced here is public-record information, not a legal opinion that a license is valid or in good standing; confirm with the issuing authority before relying on it.
  • What we cannot tell you. These sources do not give us a publication or effective date on ingest, and they state no update cadence or SLA that we capture, so neither is shown anywhere on this site. Where a signal is missing we print “not established” rather than a placeholder date.