How do you start a cannabis dispensary business in 2027?

What a dispensary actually is, and why the state you pick decides everything

A licensed dispensary is a state-regulated retail operation that sells flower, pre-rolls, vapes, edibles, concentrates, tinctures and topicals to adult-use consumers, registered medical patients, or both. It is legally distinct from cultivation (growing), processing and extraction (turning biomass into concentrates and edibles), delivery-only operations, and hemp/CBD retail — the last of which sits outside the Controlled Substances Act entirely under the 2018 Farm Bill and therefore does not carry any of the tax or banking pain described below.

The first decision, and the one that determines whether your business earns healthy margins or fights for single digits, is which state you enter. States fall into two broad regimes, and the gap between them is not marginal — it is the difference between a business model and a hobby with inventory.

Capped-license states limit how many retail licenses exist, either by statute or by regulatory rulemaking. Florida issues a small number of vertically integrated medical marijuana treatment center permits. Pennsylvania, Maryland, Ohio, Massachusetts, New Jersey and New York all run capped or slow-issuance frameworks. When supply of licenses is artificially constrained, wholesale prices stay firm, retail prices hold, and mature stores commonly land in the high-teens-to-thirty-percent EBITDA range. The catch is that the license itself becomes the asset — and its price reflects that scarcity.

Unlimited-license states issue retail licenses to anyone who meets the standards, subject only to local zoning. California, Colorado, Michigan, Oregon and Washington all work this way. Entry cost collapses to a few hundred thousand dollars. So does margin. Supply expands until price compresses, and single-store independents in these markets frequently struggle to clear high-single-digit EBITDA. Michigan and Oregon in particular have seen flower prices fall dramatically as license counts climbed into the high hundreds and beyond.

There is a third path worth naming: social-equity licensing. New York's CAURD program, Illinois' social equity applicants, Massachusetts' Economic Empowerment tier and New Jersey's equity classes all issue licenses at low or zero cost to applicants who meet criminal-justice-impact, residency and ownership criteria. The license is cheap. The constraint is transfer restrictions — multi-year holds before you can sell — plus, in several states, a regulator that has grown openly hostile to management-services arrangements that strip equity from the named licensee. If you are entering through a social-equity door, price the illiquidity honestly and structure the capital partnership so that it survives regulatory scrutiny rather than merely surviving the closing.

Federally, cannabis remains a Schedule I substance as of 2026. The DEA published a proposed rule in August 2024 to move it to Schedule III; the comment period closed, the administrative process stalled, and no final rule has landed. Rescheduling would eliminate the §280E tax penalty, open mainstream banking, and likely allow US exchange listings — which is why nearly every capital decision in the industry is currently sitting in a holding pattern waiting on it. Plan your business assuming it does not happen on your timeline. If it does, treat it as upside.

The step-by-step process from decision to opening day

The sequence matters more than people expect. Doing these out of order — signing a lease before confirming zoning, for instance, or applying without a site — is the single most common way first-time operators burn six figures before they sell anything.

Step one: choose the state and license type. Read the actual statute and the current rulemaking, not a summary. Confirm whether the state issues new licenses at all right now, whether there is a pending application window, and whether vertical integration is required, permitted, or forbidden. Florida requires vertical integration; most retail-only states forbid an unlicensed operator from touching cultivation.

Step two: secure zoning before anything else. Nearly every municipality applies setback requirements — commonly in the range of several hundred to fifteen hundred feet from schools, parks, daycares and sometimes churches. Many cities ban retail cannabis outright even in legal states; others cap the number of stores per district. Pull the municipal code, map the eligible parcels yourself, and confirm with the planning department in writing before you commit to a site. In several states you also need a host community agreement or local letter of support before the state will even score your application.

Step three: lock the site with a contingent lease or option. You almost always need site control to apply, but you cannot afford to pay rent through an 18-month approval process on a license you may not get. The standard structure is a lease with a licensing contingency and a delayed rent commencement tied to license issuance, or an option agreement with a modest non-refundable fee. Landlords who have done cannabis deals understand this; landlords who have not will need educating, and will likely price a premium for the perceived risk.

Step four: build the entity and the capital stack. C-corporation is the near-universal default, not because it is elegant, but because §280E makes pass-through structures worse — the disallowed deductions flow to individual rates rather than the corporate rate. Separate the real estate into its own entity from day one if you own it; that structure is what makes a later sale-leaseback possible without unwinding everything.

Step five: file the application. In competitive capped-license states, applications routinely run into the hundreds of pages, covering operations plans, security plans, financial capacity, diversity and community-impact commitments, standard operating procedures, and local approval evidence. Application fees range from a few thousand dollars in some programs to well over a hundred thousand in others, and the soft cost of preparing a competitive submission — attorneys, consultants, architects, subject-matter experts — often exceeds the fee itself by an order of magnitude. Approval rates in genuinely competitive windows are low.

Step six: build out and pass inspection. Vault or safe, camera coverage of every product touchpoint, alarm system with state-approved monitoring, limited-access areas, ADA compliance, an age-verification vestibule, and full seed-to-sale integration all have to be installed and demonstrably working before the state signs off.

Step seven: bank, staff, badge, stock and open. Employee badging alone can take a couple of months per person in some states, so start it early.

Costs, timelines and the ranges that actually show up in budgets

Break the budget into five buckets: license, build-out, working capital, compliance overhead, and the tax drag that quietly eats profit no other retail category has to plan for.

License cost. In unlimited-license states the state fee plus local permitting typically lands in the low-to-mid six figures once you include the consultants, the conditional-use hearing, and the legal work. In capped-license states you are usually buying an existing license on the secondary market rather than winning a new one, and that is where the numbers get serious — mid-seven figures to low-eight figures depending on state, whether the permit is vertically integrated, and how many retail locations it authorizes. Florida's vertically integrated permits and New York's registered-organization licenses sit at the top of that range; Massachusetts and California sit near the bottom. Prices moved meaningfully during the 2023–2026 downturn, and distressed sellers exist in most markets.

Build-out. A 2,500 to 5,500 square foot store runs roughly $1.5M–$3M in an unlimited-license market and $3M–$6M in a capped market where finish expectations are higher. That covers the vault or UL-rated safe, thirty-plus cameras with the state-mandated retention window (commonly 90 days, longer in some states), alarm and panic systems, limited-access construction, ADA work, millwork and fixtures, point-of-sale hardware, and the initial wholesale inventory load.

Working capital. Plan for at least six months of operating expense on top of the build. Rent, payroll, insurance and software do not wait for your ramp.

Recurring compliance costs. Seed-to-sale tracking software plus per-tag RFID fees; point-of-sale software per location plus a percentage on online orders; menu listings on the major cannabis marketplaces, which can run several thousand dollars a month for good placement in a competitive metro; batch lab testing at a few hundred dollars per batch; child-resistant packaging at well under a dollar to a couple of dollars per unit, which sounds trivial until you multiply it across every SKU you sell; armored cash transport monthly; and insurance premiums that are dramatically higher than comparable retail because carriers price robbery risk, product liability and federal illegality into every policy.

Timeline. Unlimited-license states with cooperative municipalities can go from decision to open in roughly 9–18 months. Capped-license application paths routinely take two to three years from window announcement to opening, and that assumes you win. Buying an existing license compresses the timeline to roughly 6–12 months, gated mostly by regulator approval of the change-of-ownership, which is itself a background-check and financial-disclosure process for every beneficial owner above a threshold percentage.

The tax drag. IRS §280E denies deductions and credits for businesses trafficking in Schedule I or II substances. In practice that means rent, marketing, retail payroll, professional fees, utilities, insurance, software and depreciation on non-production assets are all non-deductible federally. Only cost of goods sold reduces taxable income. Effective federal rates commonly land far above the 21% corporate baseline. The mitigation is careful §471 cost-of-goods allocation — pushing defensible costs like inbound freight, packaging, compliance testing, production-area security and production-area depreciation into COGS — which materially reduces the bill but never eliminates it. This is not a place to save money on accounting. A cannabis-specialist CPA and tax attorney are load-bearing, and the Tax Court record — *Champ*, *Olive*, *Harborside* — has repeatedly punished aggressive allocation that lacked documentation.

Where operators get it wrong

Underwriting to gross margin instead of after-§280E cash. A store doing $6M in revenue with a healthy gross margin can still generate almost no distributable cash after federal tax, because tax is assessed on gross profit rather than net income. Build your model bottom-up on after-tax cash, or you will hire and expand against money you do not have.

Treating banking as a solvable Tuesday problem. Most national banks will not take cannabis deposits. You will bank with a credit union or a specialty platform, pay monthly account fees plus deposit-based charges, and manage a large cash position. The cashless-ATM workaround that many stores relied on was shut down by the card networks in 2023; credit cards are not an option. PIN-debit and ACH-style payment providers are the current answer, and a meaningful share of transactions remain cash. Open your account before you sign a lease — being unbankable after committing to rent is a genuinely common failure.

Sloppy inventory reconciliation. Seed-to-sale variance thresholds are tight, often in the sub-one-percent range before an audit is triggered. Larger variances can move into suspension territory. The fix is boring and structural: daily counts on high-value categories, a named compliance owner, POS-to-tracking-system integration rather than manual entry, and documented destruction events. Most license actions come from paperwork failures, not from anything dramatic.

Buying a license without diligencing the entity. In a stock purchase you inherit tax liabilities, regulatory history, and any undisclosed ownership arrangements. Structure as an asset deal where the state permits it, and independently confirm the license is in good standing with the regulator rather than trusting the seller's representation.

Assuming rescheduling will bail out the model. Every distressed operator from 2022 onward assumed relief was one legislative session away. Underwrite to today's rules.

Neglecting the RevOps discipline of the storefront. Cannabis retail is retail. The operators who outperform run it like a revenue operation: they track basket size and category mix by daypart, measure budtender attach rate on accessories and pre-rolls, run loyalty programs with actual cohort analysis rather than blanket discounting, and know their inventory turn by SKU. Turning stock two to four times a year is the difference between fresh product and markdowns on expiring flower. Discount-driven traffic — the twenty-percent-off midweek promotion, the Sunday ounce deal — reliably brings people in and reliably compresses margin, and a surprising number of stores have never measured which effect is larger for them.

Understaffing compliance while overstaffing the floor. Budtenders are visible; the compliance officer is not. States including Massachusetts, New York, Illinois, New Jersey and Ohio effectively require a dedicated compliance role, and every other state rewards having one.

A decision framework for choosing your entry path

There is no universally correct answer here — there is a correct answer for your capital, your risk tolerance, and your timeline. Work through it in that order.

If you have under roughly $2M in total available capital, a capped-license state is not realistic as a de novo entry. Your viable paths are an unlimited-license market where you compete on brand, location and operations rather than on scarcity, or a social-equity application if you genuinely qualify — not if you are being recruited as a figurehead by someone else's capital, which is both ethically ugly and increasingly enforcement-exposed.

If you have $5M–$30M, the calculus shifts toward buying an existing license in a capped state, where the license premium is real but so is the margin structure that justifies it. In this range, distressed acquisitions deserve serious attention: several markets have operators who built well and were then crushed by price compression and tax drag, and their infrastructure trades far below replacement cost. You inherit a turnaround, but you skip the application lottery entirely.

If you can commit above that, the strategic question changes from "should I open a store" to "how do I build a state cluster." The economics of this industry reward density — shared back office, central purchasing, one compliance function across several locations, brand recognition within a single regulatory regime. Vertical integration, where the state allows it, adds another layer: growing your own product captures the wholesale margin you would otherwise pay away, and it legitimately expands the cost base you can allocate to COGS under §471.

Two adjacent paths deserve a mention because they solve for the same demand with different risk profiles. Delivery-only licenses, where states offer them, strip out the build-out and setback problems almost entirely — no vault-grade retail floor, no conditional-use hearing, dramatically lower capital — at the cost of thinner margins and a harder brand-building problem. And ancillary businesses — point-of-sale software, packaging, security services, testing labs, compliance consulting, real estate leased to operators — serve the same market without ever touching the plant, which means no §280E, normal banking, and normal financing. Plenty of people who set out to open a store discovered the better return was selling to the hundred stores around them.

Related questions

Can I open a dispensary in a state where cannabis is only medically legal?

Yes, but the license is typically called something else — a medical marijuana treatment center, a dispensary permit, or a registered organization — and patient counts cap your addressable market. Medical-only states often have fewer licenses and better margins, but slower revenue growth than adult-use markets.

Do I need to grow my own cannabis to open a store?

Usually no, and in many states you are prohibited from doing both. Vertical integration is required in a handful of states, optional in some, and forbidden in others. Retail-only operators buy from licensed cultivators and distributors at wholesale.

How much cash will I actually handle day to day?

More than any comparable retail category. With credit cards unavailable and a meaningful share of customers paying cash, a busy store can move tens of thousands of dollars weekly in physical currency, requiring a vault, dual-control counting procedures, and contracted armored transport.

Is buying an existing dispensary safer than applying for a new license?

Generally yes on timeline and certainty, and generally no on price. You skip a low-probability application process and inherit revenue, but you pay a license premium and assume the seller's regulatory and tax history unless you structure it carefully as an asset purchase.

What happens to my business if cannabis is rescheduled?

The §280E tax penalty would likely disappear, which is a large margin improvement, and banking and exchange listings would open up. It would also lower barriers to entry and probably accelerate competition and price compression. Do not build a model that requires it.

FAQ

What does it realistically cost to start a cannabis dispensary?

In an unlimited-license state, roughly $1.5M–$4M all-in covering license and permits, build-out, inventory and six months of working capital. In a capped-license state where you must buy an existing license, total capital commonly runs $10M–$30M, with the license itself representing the majority of that. Budget conservatively — cost overruns during regulated build-out are the norm, not the exception.

How long does licensing take?

Six to twelve months if you are acquiring an existing license and the regulator is processing change-of-ownership normally. Nine to eighteen months in unlimited-license states with a cooperative municipality. Two to three years in competitive capped-license application processes, and that assumes you win the window — many applicants do not, and reapplication means waiting for the next one.

Why can't I just use a normal bank and accept credit cards?

Because cannabis remains federally illegal, banks face Bank Secrecy Act and anti-money-laundering exposure for servicing plant-touching businesses, and the major card networks prohibit cannabis merchant category codes. Your practical options are a cannabis-friendly credit union or specialty banking platform, PIN-debit and ACH-style payment providers, and cash. Federal safe-harbor banking legislation has repeatedly stalled in Congress.

What is §280E and can I get around it?

Section 280E of the Internal Revenue Code denies deductions and credits to businesses trafficking Schedule I or II controlled substances, so a dispensary can only reduce federal taxable income by cost of goods sold. You cannot get around it, but you can reduce it through documented §471 cost allocation and, where permitted, vertical integration that legitimately moves more of the cost stack into production. Use a cannabis-specialist CPA.

What most often gets a dispensary license suspended or revoked?

Inventory tracking failures. Seed-to-sale variance beyond the state threshold triggers audits, and repeated or large discrepancies escalate to suspension proceedings. Other common causes include selling to underage or unregistered buyers, camera or retention lapses discovered during inspection, unreported ownership changes, and operating outside approved hours or premises boundaries.

Should I hire a cannabis consultant for the application?

In competitive capped-license windows, effectively yes — applications are scored, and experienced writers know what scores. In unlimited-license states with a checklist-style process, a cannabis attorney and a good architect usually suffice. Vet consultants on outcomes in the specific state you are entering, not on general industry tenure, and avoid anyone promising a guaranteed award.

Sources

1. IRS — Marijuana Industry Tax Center — official guidance on §280E application to cannabis businesses. https://www.irs.gov/businesses/small-businesses-self-employed/marijuana-industry
2. U.S. Drug Enforcement Administration — scheduling status and the 2024 proposed rescheduling rulemaking. https://www.dea.gov
3. FinCEN — BSA Expectations Regarding Marijuana-Related Businesses — federal guidance governing cannabis banking. https://www.fincen.gov
4. California Department of Cannabis Control — retail license types, application process and fees. https://cannabis.ca.gov
5. New York State Office of Cannabis Management — adult-use retail licensing and social-equity programs. https://cannabis.ny.gov
6. Colorado Marijuana Enforcement Division — retail licensing, inventory tracking and compliance rules. https://sbg.colorado.gov/med
7. Michigan Cannabis Regulatory Agency — adult-use retailer licensing and market data. https://www.michigan.gov/cra
8. Massachusetts Cannabis Control Commission — retail licensing, host community agreements and Economic Empowerment. https://masscannabiscontrol.com
9. Florida Office of Medical Marijuana Use — medical marijuana treatment center framework. https://knowthefactsmmj.com
10. MJBizDaily — industry reporting on market size, license values and dispensary economics. https://mjbizdaily.com

Related on PULSE

• [How Many Budtenders Should I Schedule Each Shift at My Cannabis Dispensary?](/knowledge/q15602)
• [How Do I Negotiate a Cannabis Dispensary Lease Without Getting Gouged?](/knowledge/q13733)
• [How Do I Negotiate a Lease and Buildout for Cannabis Cultivation?](/knowledge/q13805)
• [How'd you fix Story Cannabis' revenue issues in 2026?](/knowledge/q1224)
• [How do you start a concrete floor coating business in 2027?](/knowledge/q9739)