Cannabis business models explained

The models at a glance

A cannabis license says what you may do. A business model says how you get paid for doing it. Two holders of the same license can run very different businesses, and a buyer values the model, not the paper.

Model	Earns money from	Biggest cost	Main risk	280E burden on adult-use income
Retail store	Markup on purchased product, times traffic	Product, payroll, rent	New stores nearby; price cuts	Heaviest
Commodity cultivation	Pounds sold at the market price	Labour, power, facility	Wholesale price falls below cost	Lightest
Craft or branded cultivation	A price premium per pound or per jar	Labour, genetics, marketing	Premium erodes; small scale	Light
Manufacturing, owned brand	Branded units sold to stores	Inputs, packaging, sales team	Shelf space and category shifts	Medium
Toll or white-label processing	Fees per unit processed for others	Equipment, labour	A few customers leave	Medium
Distribution	Fees or margin for moving, storing, and collecting	Fleet, warehouse, receivables	Customers who do not pay	Heavy
Delivery	Retail margin without a storefront	Drivers, vehicles, dispatch	Low order size; local bans	Heavy
Testing laboratory	Fees per sample	Instruments, chemists, accreditation	Price wars; regulatory sanctions	Heavy, as a service business
Vertically integrated	Every margin from plant to sale	All of the above	Capital needs; complexity	Blended
Ancillary, non-plant-touching	Goods or services sold to licensees	Varies	Customer credit risk	Generally none

The rest of this guide takes each model in turn. Licence names, state-by-state rules, and sale mechanics for each are on the license types hubs.

Retail: a location business

A dispensary earns the spread between wholesale cost and shelf price, multiplied by the number of transactions. It owns no production. Its real asset is a locally approved address that competitors cannot easily copy.

Headset recorded an average basket of $47.29 and 2.74 items per basket across 16 states in the 12 months ending June 2026, with the basket down 5.5% from the prior year and discounts running at 26.0% of shelf value. Headset data reported by Cova put the median store’s gross margin at 42.7% in 2025, down from 52.6% in 2021.

What that means for the model:

• Falling wholesale prices help a retailer’s cost, but shelf prices fall too, and rent and payroll do not.
• The defence is a location protected by a local cap or buffer rules, plus loyalty and house brands.
• Retail carries the heaviest 280E burden, because a reseller’s cost of goods is essentially the wholesale invoice.

Buyers pay for verified store-level cash flow and for how hard the location is to replicate. Details are on dispensaries: how they are bought, sold, and valued and dispensary valuation multiples.

Cultivation: a commodity business, unless you can prove otherwise

A grower sells pounds. In most states flower is a commodity, so the model is won on cost per pound against the state’s wholesale price.

Public growers show how wide the range is. Grown Rogue reported indoor flower costs of $277 per pound in Michigan, $326 in Oregon, and $670 in New Jersey for the quarter ended 30 June 2026. Glass House Brands, a California greenhouse operator, reported production cost of $122 per equivalent dry pound against an average selling price of $211 for the same quarter. A $670 cost works in New Jersey because prices there are several times higher. It would not survive in Michigan.

There are two versions of the model:

1. Low-cost scale. Large canopy, greenhouse or efficient indoor, selling bulk flower and biomass. Value rests on cost position and the facility.
2. Craft or branded flower. Smaller canopy, sold in branded jars at a premium. It works only if track-and-trace and wholesale invoices show the premium is real and repeatable.

Cultivation has the lightest 280E burden because production labour, utilities, nutrients, and facility costs go into inventory. Its weakness is price. See cultivation businesses, and compare it with retail on dispensary vs cultivation as an acquisition.

Manufacturing and brands: owned demand or rented capacity

A manufacturer turns biomass into vapes, edibles, concentrates, pre-rolls, and beverages. The question that decides value is who owns the customer.

• Owned brand. The company makes and sells its own branded products. It earns a brand margin, and it needs a sales force and shelf placement. Category matters: Headset’s data for the year to June 2026 shows pre-rolls up 10.1% and concentrates down 7.7%. LeafLink’s 2026 pricing guide, based on 2025 wholesale transactions, shows cartridge prices down 12.0% and edibles up 1.7%.
• Toll processing and white label. The company processes other people’s material or makes products under their labels for a fee. Inventory risk is low. Customer concentration is high, and month-to-month work is worth little more than the equipment that performs it.
• Brand licensing, asset-light. The brand owner holds recipes, packaging, and trademarks where available, and licenses them to licensed manufacturers state by state for a royalty. Capital needs are low and the model crosses state lines, which plant-touching operations cannot. Some regulators treat royalties or profit shares as financial interests to be disclosed; see true party of interest and financial interest rules.

Manufacturers are producers for tax purposes, so inputs, production labour, and packaging sit in cost of goods sold, but sales and marketing do not. Valuation by revenue stream is on manufacturing and extraction valuation.

Distribution: logistics plus credit

A distributor moves product between licensees, stores it, arranges testing in some states, and often collects payment from retailers on behalf of brands. It earns a fee or a margin on each unit.

The model looks like logistics and behaves like lending. A distributor that extends terms to retailers is financing them. California’s Herbl reportedly handled $700 million of product in 2022 and collapsed in 2023, leaving brands unpaid. Its assets went through a receivership, not bankruptcy.

Not every state has a separate distribution tier. Where growers and manufacturers may sell directly to stores, distribution is a service some operators buy and others do in-house. Buyers look at receivables ageing, customer concentration, and fleet and warehouse compliance. See distribution businesses.

Delivery, lounges, and testing labs

Delivery is retail without the storefront. It avoids premium rent but adds drivers, vehicles, insurance, and dispatch software, and order sizes have to cover the cost of each trip. Many cities that allow stores do not allow delivery, and the reverse. See delivery businesses.

Consumption lounges sell an on-site experience where the state and city allow it. The model is young, local rules vary widely, and few have a long earnings record to value. See consumption lounges.

Testing laboratories sell a regulated service: fees per sample from growers and manufacturers who must test before sale. States generally require labs to be independent of the businesses they test. Instruments and accredited staff are the fixed costs, and the risks are price competition and sanctions for inflated potency results. See testing laboratories.

Microbusiness: a small integrated model by statute

A microbusiness licence lets one small operator grow, make, and sell under a single licence, within tight size limits. California’s Type 12 must carry out at least three of four activities and may cultivate no more than 10,000 square feet. Other states set plant counts or canopy limits and restrict who may own one.

The model captures the full margin on a small volume, much like a craft brewery with a taproom. The limits that make it accessible also cap its growth and narrow the pool of eligible buyers. See microbusinesses.

Vertical integration: owning every margin

An integrated operator grows, processes, and sells through its own stores. It keeps the wholesale margin and the retail margin, controls supply, and can fill its shelves with house brands.

In some states this is not a choice. Florida requires every medical marijuana treatment center to cultivate, process, transport, and dispense, and the state supreme court upheld that mandate in 2021. In others it is forbidden: Washington bars a producer or processor from holding a direct or indirect financial interest in a retailer. Most states sit in between.

The numbers can be strong in limited-license markets. Trulieve, which is integrated and concentrated in Florida, reported a 60% gross margin for the second quarter of 2026. The costs are capital, complexity, and the fact that a weak segment can hide inside a strong one. Buyers restate internal transfers at market wholesale prices and value each segment separately. See vertically integrated groups.

Ancillary and management-company models

Ancillary businesses sell to the industry without touching the plant: packaging, equipment, software, security, staffing, consulting, and cannabis real estate. They need no cannabis licence, are generally outside 280E, and have more normal access to banking. Their revenue depends on the health of their cannabis customers.

Two variations need care:

• Real estate holding companies. An owner holds the building in one entity and leases it to the licensed operator. Rent is steadier than cannabis profit, and the property can be sold or financed separately. See cannabis real estate basics.
• Management companies. A company supplies staff, systems, or management to a licensee for a fee. The Tax Court held in Alternative Health Care Advocates (2018) that a management company serving a dispensary was itself trafficking under 280E. State regulators may also treat a manager with control or a profit share as an owner. The model is not a way around either rule.

The investment trade-offs are compared on plant-touching vs ancillary investment, and sale issues are on ancillary cannabis businesses.

How 280E treats each model

Section 280E disallows deductions other than cost of goods sold for businesses trafficking in Schedule I or II substances. Adult-use marijuana is still Schedule I. State-licensed medical activity moved to Schedule III in April 2026 and is outside 280E going forward, so dual operators must split their results.

Because the rule turns on cost of goods sold, the model decides the tax.

Illustrative, from the comparison on 280E by license type: three adult-use businesses each earn $500,000 before tax.

Model	Federal tax under 280E	Effective rate on pre-tax profit
Retailer	$472,500	94.5%
Manufacturer	$367,500	73.5%
Cultivator	$315,000	63.0%

The assumptions behind those figures are on that page. The point is the ranking: the more of a model’s spending that counts as production cost, the more of its profit survives federal tax. An integrated operator lands in between, depending on how much profit is earned at the production level.

Which models sell best, and why

Buyers pay for earnings they can verify and for things competitors cannot copy.

Model	What a buyer is really paying for	Evidence they will want
Retail	A protected location with steady traffic	Point-of-sale tied to track-and-trace; lease; local cap
Cultivation	Cost per pound below the state price; the facility	Harvest and cost records; realised prices
Owned brand	Repeat wholesale orders at a premium	Sales by account over time; pricing against generic
Toll processing	Contracts and equipment in place	Written agreements with volume and term
Distribution	Collectible receivables and customer spread	Ageing report; concentration; write-off history
Vertical	Each segment on its own, plus license scarcity	Segment statements at market transfer prices
Ancillary	Ordinary business earnings	Customer credit quality; contracts

License scarcity sits on top of every row. In a capped state a license has value even with weak operations. In an open state it usually has little. The method by model is on cannabis valuation methods by license type.

Choosing a model as a buyer

1. Start with the state, because it decides which models are legal, which can be combined, and whether licenses are scarce.
2. Match the model to your skills: retail operations, horticulture, food or chemical manufacturing, logistics, or business services.
3. Test the model at today’s prices, then at lower prices, on an after-tax basis.
4. Check the ownership rules for the combination you have in mind before you buy the first piece.
5. Think about the exit on the way in. Models with verifiable earnings and a wide buyer pool are easier to sell later.

A broker will typically ask which model fits your background and capital before suggesting any target, and that is a sensible use of a free 20-minute intro call. This guide is general education, not legal, tax, or investment advice.