LABHemp'in Studio · Research/ EXP-01 · State of Hemp2026-11-15 20:43:30 UTCAlpha

EXP-01 · Evidence·Edition 01 · spring 2027 (proposed)

State of Hemp

The yearly measured account of hemp in the real economy: who asked for it in public, where it is legal, who supplies it, how much carbon it holds, and which claims about it survive the evidence. It forecasts nothing. Every chart is a file you can download and cite.

Live counts · 30 Oct 2026DOIs reserved at freezeEN · FR · ES
D1 · PUBLIC TENDERS NAMING HEMP
JAN 2024 → 30 OCT 2026
108 PROCEDURES
15 COUNTRIES
● hemp in lot or spec · ● relaunched
SELECT A CLUSTER · SOURCE TED
Alpha
A working alpha of the State of Hemp report. Counts tagged Sourced were read from Hempin’s own records on 30 October 2026, from public sources named under each figure. Every DOI, licence, page number and launch date is a placeholder until it is set.

00The cover finding

108 times, a European public buyer wrote hemp into a tender, in their own language.

TED, Publications Office of the EU · 15 countries · since January 2024 · measured 30 Oct 2026.

Procedures naming hemp
0
Sourced since 1 Jan 2024
In a lot or the spec
0
Sourced not a passing mention
Buyer countries
0
Sourced France wrote 55
Relaunched
0
Verify census; live table reads 9
Supplier matches
0
Sourced to 34 companies

01How to read this report

A report that shows its working

Hemp has had plenty of reports. Most are forecasts, and most of the forecasts were wrong. This one counts what can be counted from public records and Hempin’s own graph, says how each number was made, and gives you the file to check it.

§Four rules for every page

Rule 1

A number carries its tier

Measured, cited, declared or not yet known, in its colour, next to the number.

Rule 2

A chart carries its dataset

Dataset id, version and DOI under every figure.

Rule 3

Untested is printed

A claim we could not test is printed as untested, not left out.

Rule 4

Gaps close the book

The last chapter lists what this edition could not measure: the work for next year.

This edition is deliberately thin in places. Stored carbon starts at zero recorded buildings. We would rather publish a small true number than a big borrowed one.

◐Every figure is tiered

Measured

Counted by Hempin from a public or own record, with the query published.

Cited

From a named external source we opened and read.

Declared

Stated by a company or grower, not independently checked.

Not yet known

Printed as a gap, with what it would take to fill it.

02Chapter 01 · Demand

Public money is already asking for hemp. It often gets no usable answer.

The headline chapter, and the one the press date is built around. Pick a country here or on the map above: both follow your choice.

▤Figure 1.1 · By buyer country

Dataset D1 · v1.0 · 108 rows · doi:10.5281/zenodo.(reserved at freeze) · Source: TED, Publications Office of the EU

◎Pick a country

Sourced

Between 1 January 2024 and 30 October 2026, 108 procedures on TED, the EU’s tender journal, named hemp in text the buyer wrote: chanvre, Hanf, hennep, cânepă, коноп. In 81 of them hemp is in a lot title or the specification.

France wrote 55. The other 53 procedures are a surprising mix: Dutch fibre-crop pilots, an Irish decorticator, eight tonnes of carded hemp for an Italian yarn tender, and plumber’s hemp for sealing pipes in Romanian and Hungarian maintenance contracts.

Select a bar, or a cluster on the map above.

▤Figure 1.2 · What the notices ask for

Sourced
Construction68
Insulation18
Textile, rope14
Other14
Growing, agri9
Research, tests8
Mulch, bedding6
Food4

A notice can ask for several families, so these add up to more than 108.

D1 · field families · buyer text first, CPV second · doi:10.5281/zenodo.(reserved at freeze)

▥Figure 1.3 · By year first published

Sourced
010203040382024372025332026** to 30 Oct, a partial year
Read with care
A flat line is not a trend. Three years is too short to say demand is rising or falling, and 2026 is a partial year. Edition 02 adds BOAMP, France’s below-threshold feed, which will raise French counts and break comparability. Both series will be printed.
D1 · field first_published_at · 2026 counts to 30 Oct · doi:10.5281/zenodo.(reserved at freeze)

↻The relaunch file

Sourced
BuyerProjectFirst seenWhat was relaunched
Conseil départemental de la ViennePôle socio-éducatif et sportif, MirebeauDec 2024Lot 03, béton de chanvre
SEMCODA for the Commune of VilleurbanneGroupe scolaire Croix Luizet2024–2025Lots 02A, 02B, 05; the first call closed 6 lots without a winner
SPL Territoires 70Former collège Gérôme, VesoulDec 2025Lots declared infructueux, hemp insulation and works
Mairie de Ville-la-GrandEspace éducatif de CornièresJun 2025Construction lots
Ville de LoosCité des enfants et des parentsMar–Jun 2025Lots relaunched twice after the 14-lot call
Commune de Labastide-Saint-SerninNew groupe scolaireJun 2024Construction lots
Reconcile
The dry-run census of 30 Oct 2026 counted 8 relaunches; the live table of the same day holds 9, because one procedure is folded differently. The edition prints whichever number survives the reconciliation, with the reason.
D1 · filter relaunch = true · six shown; the full list is in the dataset · doi:10.5281/zenodo.(reserved at freeze)

?What a relaunch proves

Under France’s public procurement code a lot is declared infructueux on one of three grounds. Pick one:

No bids

Nobody answered the lot at all.

It proves

A public body specified hemp, published it, and did not get an answer it could use.

It does not prove

That no hemp supplier exists.

Twenty procedures closed at least one lot without a winner. For each relaunch, edition 01 reads the notice and prints which ground applied, where the notice says.

Since then
Since 30 October 2026 every new hemp tender is ingested daily and matched to companies on the record: 848 matches so far, to 34 companies. Edition 02 reports how many of those companies bid, and how many won.

03Chapter 02 · Legality

237 territories, 1,820 current facts, every one with a source.

The Atlas as a yearly snapshot with a DOI, so a regulator or a lawyer can cite the state of the map on a given day. On 30 Oct 2026 none of the 1,820 current facts was past its review date.

▥Figure 2.1 · Overall status

■ Legal
67
■ Restricted
18
■ Prohibited
111
■ Unknown
1

Of 237 territories in the Atlas, 197 carry an overall status. “Legal” is a summary of many rules: the status chart is the most quoted kind of figure in this industry, and the easiest to misread.

Dataset D2 · Atlas legality snapshot · field overall_status · doi:10.5281/zenodo.(reserved at freeze) · definitions in the methods appendix

▤Figure 2.2 · What the facts are about

Sourced
Cultivation395
CBD345
Cannabis313
Status197
Legislation145
Trade99
Narrative99
Industrial57
D2 · field category · top categories · doi:10.5281/zenodo.(reserved at freeze)
Before print
The status values are Hempin’s own summary field. The edition prints its definition beside the chart and spot-checks the ten largest producers against the law they cite. Breadth is not depth: the Atlas does not yet cover the licensed operator base of any country.

04Chapter 03 · Supply

Pages versus people

A record is only shared when the company stands behind it. A claimed page is the difference between a record and a scrape, so the report prints the claim count first, every year, whatever it is.

⇣Figure 3.1 · From public demand to a company that answered

Sourced
Tenders naming hempD1108Supplier matches madematcher v1848Companies matchedof 636 active records34Matched companies that bidnot yet trackedblank until edition 02Matched companies that wonnot yet trackedblank until edition 02bar length ∝ √n

Bids and wins are blank on purpose. Award notices name winners, but some are sole traders, so winners are stored only once a lawful basis is written down. Edition 02 fills these bars or explains why not.

D1 + D3 · registry snapshot, public fields only · doi:10.5281/zenodo.(reserved at freeze)

▦636 records · 9 claimed

Sourced

Each square is an active company page anyone can read and the company can claim. The lit squares are the ones a company stands behind. The edition prints the count on its freeze date, beside last year’s, with no rounding.

636 active records in 43 countries on 30 Oct 2026 · 9 claimed on 31 Aug 2026 · positions are illustrative

05Chapter 04 · Stored carbon

Absorbed is not stored. We count carbon in walls, not in fields.

The chapter that starts empty, by design. Most of what a field draws down returns to the air within a few years unless the plant becomes something that lasts.

◷The Hemp Clock at freeze

No data
0.0t CO₂ recorded

0 buildings recorded with mix, density, area and method. Target for edition 02: the first 10 corridor buildings in France.

Dataset D5 · v1.0 ships as an empty, documented schema.

↧Absorbed in the field

Verify
0 t5 t10 t15 t20 tPublished estimates7–15 tThe figure people quote22 tt CO₂ per hectare, absorbed in biomass · absorbed is not stored

A hectare of hemp draws down roughly 7 to 15 tonnes of CO₂ into its biomass in published estimates. The number that matters is what stays locked in a building or a long-lived product, for how long, and in which mix.

Published estimates, from search summaries: re-read before print

▦Build a wall record

Sample wall

Tick what the record holds. A wall goes on the record only with all five, and only if it is meant to last.

Mix
Area420 m²
Life—

≡What the record would say

0 · neutralnet emitted ←→ net storednot on the record

Not on the record. Missing 3: density, kg/m³, method and source, expected life.

The same study measured both: a light, lime-led hempcrete wall stores up to 16 kg CO₂ per m²; a cement-heavy mix at 425 kg/m³ still emitted 8.16 kg per m². That is why every wall carries its mix, density and binder, or it does not go on the record.

Never a credit
No carbon credit is described or implied. Verra’s hemp methodology has been on hold since December 2023, with a reassessment planned for Q3 2027.
Per-m² figures: Arehart et al. 2020, via Carbon Leadership Forum · the wall itself is a sample

06Chapter 05 · Claims tested

Eight claims people repeat, graded against what was measured

Hemp’s last boom died partly of its own hype. Grades can move up or down between editions; each change is printed. This chapter is the Evidence Ledger, printed.

✓The scorecard

Claim as usually toldGradeWhat was measuredMain source
Hempcrete walls store more carbon than they emitStrongUp to 16 kg CO₂/m² stored net; a cement-heavy mix emitted 8.16 kg/m². True when the mix is right.Arehart et al. 2020
Hemp textiles are as good as linenMediumComparable to linen per the European Commission; spinning capacity outside China is still experimental.European Commission
Hemp uses far less water than cottonMediumWater-footprint studies put hemp textile well below cotton; most are older and count water only.Water-footprint studies *
A hectare of hemp captures 22 t of CO₂WeakPublished estimates are 7 to 15 t absorbed in biomass; absorbed is not stored.Published estimates *
Hemp carbon credits pay farmers todayNot supportedNot yet. No hemp credits are issued; Verra’s method has been on hold since Dec 2023, review Q3 2027. EU CRCF covers 35-year product storage.Verra; EU CRCF summary
Hemp improves the next cropWeakOne six-year Catalonia study saw wheat yields rise after hemp; few rigorous trials.Catalonia rotation study *
Hemp cleans polluted soilWeakAbout 0.23 g Cd and 3.8 g Pb removed per ha per season; called “not an efficient plant” for it.Toxics field study (PMC)
Hemp batteries beat lithiumNot supportedNo hemp battery or supercapacitor has reached a product on sale.Hempin scan
Cited · Dataset D4 · one row per claim, sources and last-checked date · doi:10.5281/zenodo.(reserved at freeze) · rows marked * come from search summaries and are re-read before print

07Chapter 06 · What we could not measure

The gaps, and who could close them

The most useful page for researchers and funders. A gap is printed where the answer would be, with what it would take to fill it.

∅Gap register

Not yet known
GapWhy it mattersWhat would close itWho could
Hemp area outside the EU, on one definitionEvery global acreage figure in circulation mixes fibre, grain and flowerNational statistics read field by field into one schemaStatistics offices, FAO; Hempin compiles
Observed, not declared, fieldsLenders will not fund a hub on declared acreageA satellite pilot in one region compared to declared fieldsA remote-sensing research partner
Who wins hemp tendersTells suppliers whether bidding paysAward notices stored for companies, with a written lawful basisHempin, edition 02
Standardised hempcrete LCAToday’s ranges span net storage to net emissionOne method applied to certified productsUniversities, EPD programme operators
Storage permanence in real buildingsThe reason carbon registries put hemp on holdSensors in corridor buildings over several yearsA living buildings lab
PricesHempin has no price data and does not imply anyOut of scope for this reportPrice reporters

08Open data

Every chart is a file

How a figure in the report links to a downloadable, citable dataset in a Hempin collection on Zenodo (proposed).

⌗Anatomy of a figure footer

MEASURED. The tier, in its colour, so the reader knows how hard the number is.

Each dataset ships with a README (fields, filters, known errors), the query that produced it and a checksum. The report cites version DOIs; hempin.org cites the concept DOI.

⛁Five open datasets

Row counts sourced
D1 · contains

Buyer institution, country, NUTS, CPV, strength, families, status, relaunch, short quotes, link to notice

Never contains

Contact persons; winners who are individuals

Licence

EU reuse, Decision 2011/833/EU, with attribution

Rows at the mock-up date, 30 Oct 2026 · bar length on a log scale · licences are open decisions

09Launch

The press date the industry plans around

A proposed calendar. Dates are suggestions until they are set.

Jan 2027

Add BOAMP

Read the grounds of each relaunch.

1 Mar 2027

Data freeze

DOIs reserved; claim count taken.

March

Reconcile

Counts reconciled; Atlas spot-checked; an outside reader for chapter 5.

Early April

FR and ES

Translated editions; embargoed pack to trade and EU press.

Late April

Publish

Report, datasets and Myth versus Measured ep. 10 on one day.

Every year

Same week

Same structure; the prior edition stays online.

State of Hemp · preview108

public tenders asked for hemp in Europe since 2024.

Source: TED · Hempin, 30 Oct 2026
State of Hemp · preview8

were relaunched after the buyer got no usable answer.

Source: TED · Hempin, 30 Oct 2026
State of Hemp · preview55

of them came from France, mostly schools asking for hempcrete.

Source: TED · Hempin, 30 Oct 2026

10Part two · Myth versus Measured

The show that tests its own industry, and is trusted because of it

One claim per episode: the number people quote, then what was actually measured, then a verdict stamp. It ends on the source, on screen, long enough to photograph. Curious, never smug; when hemp wins, the show says so just as plainly.

MeasuredStrong

Repeated studies agree; the claim holds as usually told.

PartlyMedium

True under conditions the claim leaves out.

ThinWeak

Little or local evidence; the claim outruns it.

MythNot supported

Measured and failed, or never reached a real product.

0:00–0:06

The claim

Said as people say it, in a real quote or headline, source blurred.

0:06–0:15

The number

The figure that gets repeated, big on screen.

0:15–0:45

The measurement

Who measured what, where, how. Field, lab or document footage.

0:45–0:55

The verdict

The stamp lands. One line on why.

0:55–1:10

What is true instead

The honest version, which is usually still good news.

1:10–1:20

Source card

Study, year, link, last checked. Held for 4 seconds.

▶Season 1 slate

  • EP 01Hemp cleans polluted soil
    Pilot. The Slovenian field study, grams per hectare.
    Weak
  • EP 02A hectare captures 22 tonnes of CO₂
    Where the 22 came from, and why absorbed is not stored.
    Weak
  • EP 03Hempcrete is carbon negative
    Yes, with the mix on screen: the same wall, two binders.
    Strong
  • EP 04Hemp batteries beat lithium
    One 2013 paper, two parked patents, no product on sale.
    Not supported
  • EP 05Hemp uses less water than cotton
    True, from older studies that count water only.
    Medium
  • EP 06Hemp credits pay farmers now
    Not yet: the hold, the 2027 review, the 35-year rule.
    Not supported
  • EP 07Hemp improves the next crop
    One six-year trial, and why one is not enough.
    Weak
  • EP 08Hemp can replace plastic in cars
    Real in door panels; small volumes; flax leads today.
    Medium
  • EP 09Hemp is as good as linen
    Close cousins in the field, far apart in the spinning mill.
    Medium
  • EP 10Nobody wants to buy hemp
    Report-day episode: 108 public tenders, read aloud.
    Not supported
Grades follow the Evidence Ledger on the day of release

◆EP 01 · treatment

Hemp cleans polluted soil

Weak

Pilot. The Slovenian field study, grams per hectare.

60 to 90 seconds · 16:9, a 9:16 social cut is a reframe made after · EN, FR, ES voice and subtitles · links its Evidence Ledger page and its dataset row.

▣Pilot storyboard · EP 01 · “Hemp cleans polluted soil”

Weak
0:301:20
FRAME 4/616:90.23 g cadmium · 3.8 g lead
PICTURE

Numbers build one by one, per hectare, per season. Scale object: a paperclip weighs about a gram.

VOICE

About a quarter of a gram of cadmium. Under four grams of lead. The authors called hemp “not an efficient plant” for the job.

Source: Toxics field study (PMC) · last checked on the freeze date

The verdict never outruns the Ledger

An episode cannot grade a claim higher or lower than its Evidence Ledger row on release day.

Two sources on screen

At least two, one of them primary. Search summaries never appear on screen.

Claims, not companies

No episode names a company or person as the source of a myth. We test ideas.

No money talk

No prices, revenues, valuations or credit values. Carbon is data, never a credit.

Corrections are visible

If a grade changes, the episode gets an on-screen correction card and a dated note, never a silent re-upload.

Good news gets the same rigour

A strong grade carries its conditions, such as the hempcrete mix, every time.

11Sources and status

What is live, what is cited, what is not real yet

Live

Sourced

Hempin’s records, 30 Oct 2026: tenders (108 procedures by country, family, year, status, relaunch), matches (848 to 34 companies), Atlas (237 territories, 1,820 current facts, 197 with a status), active company records (636 in 43 countries). Claimed pages: 9 on 31 Aug 2026.

Cited

Sourced

Arehart et al. 2020 via Carbon Leadership Forum; the Slovenian heavy-metal field study (Toxics, PMC); Verra’s hold and Q3 2027 reassessment; EU CRCF 35-year product storage; European Commission hemp pages.

To re-read

Verify

Still from search summaries: 7–15 t CO₂/ha, the Catalonia rotation study, automotive volumes, water-footprint comparisons. Re-read before print.

Not real yet

Open question

Every DOI, the licences, the page numbers and the launch dates. The relaunch count (8 or 9) is reconciled before freeze.