During 2025, more than seven thousand Thai cannabis shops closed. That was not a market correction. It was a regulatory reclassification working through an industry that had built itself on a different set of rules. If you are thinking of putting capital into this sector, you need to understand what happened, why it happened, and what a shop’s cost structure looks like now that the model has changed. This article sets out the verified numbers, the cost lines you have to fill in with local quotes, and the risks that the closure data makes concrete.
Status as of October 2026
The closure data below runs to the end of 2025. Since then, three things have happened that matter directly to an investor:
- June 2026: a crackdown on blank (pre-signed) PT 33 prescriptions. See the PT 33 enforcement article.
- July 2026: new advertising criteria. See the advertising criteria article.
- 22 September 2026: the Cabinet approved the Cannabis Control Act and sent it to the House. It is still a draft, not law. Press reports say it would add an explicit under-20 sales ban. See the Cannabis Control Act draft tracker.
The 2026 licensing regulations did not change the PT 33 prescription rules.
1. The numbers, verified
These are the figures I could verify. They are the starting point for any analysis:
| Figure | Value | As at |
|---|---|---|
| Cannabis shops nationwide | 18,433 | 28 December 2025 |
| Shops whose licences expired during 2025 | 8,636 | 28 December 2025 |
| Of those, licences renewed | 1,339 (15.5%) | 28 December 2025 |
| Of those, shops closed | 7,297 | 28 December 2025 |
| Shops remaining operational | 11,136 | Reported 1 February 2026 |
Source: Ministry of Public Health data as at 28 December 2025, reported by The Nation (Bangkok), “Over 7,000 cannabis shops close due to stricter regulations and new law”, 1 February 2026.
A 15.5% renewal rate is the number to sit with. About 84.5% of the licences up for renewal in 2025 were not renewed. That is roughly six in seven, and about 40% of all shops.
The regulatory cause. On 9 June 2022 cannabis was removed from the narcotics list. Since then the framework has moved to medical use only. The decisive step was the Notification of the Ministry of Public Health Re: Controlled Herbs (Cannabis) B.E. 2568, in force from 26 June 2025. Under it:
- flower may be dispensed only against a PT 33 prescription from an authorised practitioner;
- a prescription covers no more than 30 days and is single use;
- flower that is sold must come from a GACP-certified site;
- online sales and advertising are banned.
The Nation describes having a licensed or registered traditional medicine practitioner on site as one of the hardest new requirements for shops. That requirement came with the June 2025 framework, which is why it could cause closures during 2025.
The financial consequence. The Nation quotes experts estimating the closures at “tens of millions of baht in investment”. That looks far too low. The same article quotes one Bangkok owner with nearly 1 million baht invested in renovations, equipment and stock. Even a few hundred thousand baht per shop, across 7,297 shops, would add up to billions. I could not find a sourced total, so treat the total loss as unknown.
Expert comment, from the same report: Dr Bundit Sornpaisarn of the Centre for Addiction and Mental Health (CAMH) in Toronto (The Nation spells the name “Bandit Sorpaisan”) calls Thailand’s policy an example of decriminalisation without a clear regulatory framework.
2. What the shakeout actually tells an investor
The regulatory risk is not theoretical. 7,297 businesses found this out in a single year. A business can be legal, operating and profitable, and still be made non-viable by a rule change with a short implementation window. Any investment model for this sector that has no scenario where the rules change again is incomplete.
The barrier to entry has risen sharply, and that cuts both ways. Practitioner involvement is a real operational constraint. It is a recruitment problem and a fixed cost, not a form to file, and it kills the low-capital shop model. It also means the remaining 11,136 operators face much less new competition than the previous cohort did.
The market is now medical, which changes the customer and the product. A shop selling to walk-in tourists and a shop dispensing against prescriptions are different businesses. They differ in margins, inventory, staffing and compliance load. A model built for the first will not survive as the second.
Consolidation may or may not be happening. 11,136 shops for about 70 million people is still a very high density. Were the 2025 closures the end of the correction or the beginning? I cannot answer that from the available data, and anyone who answers it confidently either way is guessing.
3. The cost model: structure, not numbers
I am not publishing baht figures for rent, fit-out, licence fees, salaries or inventory, because I could not verify current figures. Plausible-looking invented numbers would be the most harmful thing an investment article could contain. The only sourced data point I have is a single case: one Bangkok owner reported nearly 1 million baht invested (The Nation, 1 February 2026). That is an anecdote, not a benchmark.
What follows is the structure: what belongs in the model, and what is commonly left out.
Setup costs
| Line | Notes |
|---|---|
| Premises deposit and advance rent | Typically several months. Location drives this more than any other variable |
| Fit-out and renovation | Includes anything the licensing conditions require: dispensing area, secure storage, record-keeping space |
| Licence application | Government fees, plus professional fees if you use an adviser |
| Security and access control | Usually a licensing condition, not an option |
| Point-of-sale and record-keeping system | Prescription dispensing needs patient and product records. This is a compliance system, not a till |
| Initial inventory | Flower must come from a GACP-certified site (see section 4) |
| Signage and branding | Flower advertising is banned (see section 5) |
| Professional fees | Legal, accounting, licensing |
Running costs
| Line | Notes |
|---|---|
| Rent | |
| Practitioner salary | The line that changed the business model. Confirm exactly what your licence conditions require during opening hours. Longer hours may mean more than one practitioner |
| Other staff | |
| Inventory purchases | |
| Testing costs | Where required for the products you carry |
| Compliance and record-keeping | Staff time, systems, audit preparation |
| Utilities, insurance, banking | |
| Licence renewal | Includes the cost of the renewal process itself. The 2025 data suggests this is where businesses fail |
The lines most often missed
Practitioner cover. One practitioner cannot cover all operating hours forever: there are holidays, illness and resignations. If your model depends on one person, you have a single point of failure that closes the shop. Budget for cover, or plan shorter hours.
Working capital through a licence renewal. The 2025 data is the argument here. Model a scenario where renewal is delayed and you cannot trade for a period.
Inventory obsolescence. Cannabis flower has a finite shelf life, and a prescription-only market moves stock more slowly.
The cost of exit. Fit-out is a sunk cost, but lease obligations continue. Closing shops can stay liable for rent on premises they can no longer use.
4. The supply-side requirement that shapes the business
A medical framework pushes compliance obligations upstream into cultivation. Flower sold or exported must come from a GACP-certified cultivation site. That means documented cultivation, batch records, testing, and traceability from plant to package.
Two consequences for an investment case:
Vertical integration becomes more attractive than it was under the recreational model, because control of a compliant supply chain is worth more when compliance is the binding constraint.
Supply may be a better position than retail, but that is not proven. Retail carries the premises cost, the practitioner cost, footfall risk and licence-renewal risk. Cultivation avoids footfall risk and the retail practitioner requirement. It still carries land and facility costs, and licence and GACP-renewal risk. No comparable farm closure data has been published, so the 2025 figures describe retail only. They do not show that cultivation is safer.
5. Risks, named specifically
Regulatory change. Ranked first because the record is clear:
- 9 June 2022: cannabis removed from the narcotics list.
- 11 November 2022: flower designated a controlled herb (MoPH notification B.E. 2565).
- 26 June 2025: replaced by the B.E. 2568 notification, with prescription-only supply on PT 33 and no online sales or advertising.
- 26 April 2026: Ministerial Regulation on cannabis/hemp extracts B.E. 2569 tightens licensing.
- 30 April 2026: Ministerial Regulation on controlled-herb licences (No. 2) B.E. 2569 tightens it further.
That is three formal rule changes in the ten months from June 2025 to April 2026, before counting the 2026 measures listed under “Status as of October 2026” and the draft Act. Assume more.
Licence renewal. A 15.5% renewal rate. Whatever the causes, this is the observed point of failure.
Practitioner availability. A hard requirement that depends on a labour market to supply it. If supply is short, salaries rise and small operators are squeezed.
Advertising. Flower advertising is banned, and the July 2026 criteria apply as well. A shop that cannot advertise depends on location, referrals and repeat custom. That changes both the marketing budget and the value of the lease.
Banking and payments. Access to financial services for cannabis businesses is uncertain. Check with your bank before committing.
Foreign ownership. The B.E. 2569 extract regulation sets nationality conditions on who may hold a licence. For the exact wording, see the B.E. 2569 licensing article. If you are a foreign investor, have a Thai lawyer check your intended structure before you do anything else.
Market saturation. 11,136 shops remain. Location analysis matters more than a business plan.
6. How to build the model
- Confirm the current legal position first with a Thai lawyer, for your specific activity and ownership structure. Everything else is wasted work if this fails.
- Confirm the practitioner requirement for your intended opening hours, and price the recruitment before you sign a lease.
- Get three real quotes for every setup line. Fill in the tables in section 3 with numbers from suppliers, not from an article.
- Model three scenarios, not one: a base case, a case where the rules tighten again (the draft Act is pending), and a case where you cannot renew.
- Calculate the break-even monthly revenue. Then work backwards to the customer numbers and basket size that requires, and ask whether the location supports it.
- Model the exit. What can you recover if this does not work?
On ROI figures. I am not publishing any, and I would treat any you are shown with scepticism. A projection that does not show its regulatory assumptions is not a projection.
7. Where the opportunity may actually be
Three observations that follow from the data rather than from optimism:
Compliance services. Thousands of operators need practitioner staffing, record-keeping systems, licence-renewal support and audit preparation. The regulatory shift created that demand directly, and serving it does not require a premises licence.
Compliant supply. About 84.5% of licences up for renewal in 2025 were not renewed (≈40% of all shops). No comparable figure has been published for cultivation, and a medical framework raises the value of documented production. The risks in section 4 still apply.
Growshop and input supply. This is the picks-and-shovels position. It serves both licensed cultivation and the wider horticulture market, and it is not subject to the cannabis retail licensing regime in the same way. Still, verify this for your specific product range: inputs sold specifically for cannabis cultivation may come with their own rules.
For anyone who went through the 2025 renewal cycle: what actually stopped the shops that closed? The reported cause is the practitioner requirement. But “could not find a practitioner”, “could not afford one”, “decided it was not worth it” and “could not meet the premises rules” are four different failures, with four different lessons for anyone modelling an entry now. The people who lived through it know which it was, and that is not in any published dataset.
Related reading: the two B.E. 2569 Ministerial Regulations, the Cannabis Control Act draft tracker, shop evaluation criteria, GACP farms.
This article is general business and market information, not investment, financial, legal or tax advice. The regulatory position has changed repeatedly and may have changed again since publication, so verify the current position with a qualified Thai lawyer before committing capital. Closure figures are Ministry of Public Health data as reported by The Nation on 1 February 2026. I have deliberately not published cost or ROI estimates that I could not verify.