There is no institutional premium series for cannabis insurance, and the forecasts that fill the gap disagree with each other. What is measurable says something the growth reports miss: the number of licensed businesses has fallen 13% in two years while property rates rose up to 40%.
On 14 August 2026, Palomar Excess & Surplus Insurance went to the US District Court for the Northern District of Illinois and asked a judge to confirm that it does not have to defend one of its own policyholders.
The policyholder is Cresco Labs, one of the largest cannabis companies in the United States. The demand Palomar wants out of runs to several million dollars.
Its argument is not that cannabis is illegal, or risky, or badly run. It is narrower and much more awkward: the class actions against Cresco do not allege bodily injury or property damage arising from an occurrence, which is what the general liability and products policies are written to respond to. The claim is that the company marketed a product deceptively. On Palomar’s reading, that is simply not what it sold cover for.
Eight days later, that dispute is unresolved, and it is the most useful thing published about cannabis insurance this year.
Start with what cannot be known, because most writing on this subject skips it.
There is no institutional premium series for cannabis insurance. No statutory supplement, no rating bureau, no trade body publishing written premium the way IUMI does for marine or NCCI does for workers’ compensation. What circulates instead are market-research forecasts, and they do not agree with each other. One report puts the global market at $2.4bn in 2025 heading for $7.2bn by 2033. Another has it reaching $6.7bn by 2032 on 14% annual growth. Same category, near-identical horizons, and the two numbers describe different markets.
Neither figure is checkable, so neither belongs in a business case.
Even the dataset that should answer the question cannot. The fifteen US surplus lines stamping offices processed $90.3bn of premium in 2025, up 7.8%, and they publish the split by line: property, liability, auto, professional. Cannabis is in there somewhere, spread invisibly across those categories, with no line item of its own. The market that actually writes this class does not know how much of it there is either.
What is measurable is the industry being insured, and it is measured very well indeed, because tax authorities count it. US states have collected more than $28.4bn in adult-use cannabis tax revenue to date, with 2025 setting a record at $4.57bn. Four states did most of the collecting.
Four states collected most of the 2025 cannabis tax take
State adult-use cannabis tax revenue, 2025
Marijuana Policy Project, July 2026, compiling state revenue reporting
Those are tax receipts, not estimates. An industry that hands over a billion dollars in a single state is not a speculative one.
Here is where the vendor narrative stops matching the ledger.
Active cannabis business licences in the United States fell about 1% in the first quarter of 2026, to 36,169. That continues a slide that started at the end of 2022. Over two years the total is down 13%. Cultivation and retail together account for more than 70% of what remains.
Insurance pricing went the other way, hard.
Writing in IA Magazine in April 2026, Olivia Overman collected the numbers from people actually underwriting the class. Property for indoor growing operations was up 25% to 40%. Commercial auto for cannabis fleets was running at two to three times its 2024 level.
Premiums are rising into a shrinking pool of insureds
Change in cannabis insurance pricing against change in active licences
Pricing from Burns & Wilcox, Cannasure and Jencap via IA Magazine, April 2026. Licence counts from CRB Monitor, first quarter 2026.
A shrinking pool of insureds paying sharply more is the signature of a market that has stopped selling growth and started paying claims.
Julia Merritt, a cannabis commercial underwriter at Burns & Wilcox, put it in one sentence: “Pricing is more closely related to real loss experience rather than future growth expectations.”
That sentence is the whole story of the last two years, and it is worth reading twice, because it is also an admission. For most of the previous decade this class was priced on a projection. The projection was wrong, the losses arrived, and the rate had to move a long way to catch up with them.
Whitney Economics, which has surveyed cannabis operators for years and is the closest thing this sector has to an independent statistics office, puts after-tax profitability at 24.4% of US operators. Across all US small businesses the figure is 65.3%.
Three quarters of the insureds lose money
Share of businesses profitable after tax
Whitney Economics operator survey. The small-business baseline is Whitney’s own comparison figure.
The same firm recorded the first year-over-year revenue decline in the legal market’s history in 2025, and forecasts 2026 at $30.5bn, up 4.9%. Growth resumes, from a smaller base, split among fewer licence holders.
For an underwriter these are not background statistics. An insured that loses money is a collection risk on the premium and a moral hazard on the claim. Three quarters of this class fits that description, and the 280E tax treatment that Schedule III would lift is a large part of why.
The perils are unglamorous and mostly electrical.
Indoor cultivation means dense lighting loads running long hours in converted buildings, and electrical overload fires are the result. Lee Woodruff, who runs the cannabis practice at Jencap Insurance Services, said underwriting has tightened “most notably for cultivation risks that aren’t using LED lighting or extraction risks using volatile solvents”.
Extraction is the other one. Pulling oil from plant material with butane or propane is a chemical process being run, in some cases, by operators whose expertise is horticulture.
Then the ordinary commercial exposures, made worse by the cash problem: burglary and vandalism against premises known to hold both product and currency, and spoilage claims when equipment fails on a crop that has months of inputs already sunk into it.
The liability side is newer, less settled, and where the Palomar dispute came from.
On 4 May 2026 more than forty named plaintiffs filed a 320-page complaint in the Northern District of Illinois against Cresco Labs, Green Thumb Industries and Verano Holdings, with a companion case against Curaleaf Holdings in Connecticut. The claims span twelve states and allege that the companies marketed cannabis as therapeutic while knowing, or having reason to know, that their products carried serious health risks. They are pleaded under federal RICO, which carries treble damages, alongside state consumer protection statutes and warranty claims. The complaint draws the comparison to the Big Tobacco litigation of the 1990s itself.
Writing in Claims Journal two weeks later, Ian Stewart, who co-chairs the cannabis law practice at Wilson Elser, said the case “signals a new era of litigation risk that could reshape how cannabis companies are underwritten and insured”. His advice to underwriters was to scrutinise marketing practices with the same attention they already give to product testing and labelling.
Nothing in a crop or a grow room is doing that damage. A sentence on a package is.
Kieran O’Rourke, director of underwriting at Cannasure Insurance Services, described what carriers now do with all this: it “allows them to draw a sharp line between operators with engineered, disciplined controls and those still relying on legacy processes”.
That line is the entire underwriting proposition. Two dispensaries with identical revenue and identical square footage are not the same risk, and the difference is documentation.
The federal position, which is the reason all of this is written in the surplus lines market, shifted in 2026 for the first time in decades.
In April, Acting Attorney General Todd Blanche signed an order moving state-licensed medical cannabis and FDA-approved cannabis products from Schedule I to Schedule III, published in the Federal Register on 28 April. The same order sent the broader question, whether cannabis generally should be rescheduled, into fresh administrative proceedings. Those hearings ran from 29 June to 15 July. Post-hearing briefs, then a recommended decision, then a ruling by the DEA Administrator: unresolved as this is written.
Adult-use cannabis therefore remains federally Schedule I while state-licensed medical cannabis does not. Underwriters are being asked to price a class where the legal status of the insured depends on which licence they hold.
Rescheduling matters mostly for tax. Schedule III removes the application of Section 280E, which currently denies ordinary business deductions to companies trafficking in controlled substances. It does not fix banking, which still needs the SAFE Banking Act or something like it, and it does not by itself make an insurer comfortable.
Which is why the more interesting development was legislative. In the week of 21 July, Senators Kevin Cramer of North Dakota and Rubén Gallego of Arizona introduced the Clarifying Law Around Insurance of Marijuana Act. The CLAIM Act would bar federal regulators from penalising an insurer solely for covering a cannabis business licensed under state, local or tribal law, extend that protection to the insurer’s officers and employees, and require the Government Accountability Office to study the barriers facing minority- and women-owned cannabis operators.
“The CLAIM Act makes sure insurers can offer their commercial products to regulated marijuana-related businesses without fear of federal prosecution,” Cramer said. Gallego framed it as cannabis businesses deserving “the same access to insurance coverage as any other business”.
Cannabis is legal in some form in 41 states. A bill was still needed in 2026 to tell insurers they would not be prosecuted for writing it.
The National Association of Insurance Commissioners has run a standing Cannabis Insurance Working Group for years and publishes a periodically updated white paper on the subject. Its charge is revealing: the group exists partly to encourage the development and expansion of admitted market insurers so that coverage is adequate in states where cannabis is legal.
Regulators do not usually have to lobby for capacity in a profitable class.
California went further and wrote the absence into its rulebook. Its 2025 Export List, the register of risks a broker may take straight to the surplus lines market without first canvassing admitted carriers, added commercial cannabis operations. An export listing is a regulator formally recording that the admitted market does not write a class in any usable volume. It is the closest thing to an official statistic this line has.
The NAIC also documents the gaps that remain, and they cluster at the edges rather than the centre: smaller operators, ancillary service businesses, cannabis-infused products, social consumption lounges. And it records a detail that says more about this line than any forecast: most cannabis businesses still pay their premiums in cash, because their banks will not process the payment.
For anyone building the systems behind this, that combination is the actual specification. A class with no credible external benchmark means rating has to be built from your own loss experience and revised as it accumulates, not filed once against an industry table that does not exist. A federal position that moved twice in four months means product and eligibility rules have to change on the timescale of a Federal Register notice. Underwriting that turns on whether a grow room uses LED lighting or a solvent is a structured-data problem before it is a pricing one, and cash premium payment is a reconciliation problem most billing systems quietly assume away. Specialty classes are where platform assumptions get found out.
The market-research reports will keep forecasting a $7bn line. The licence registries say the number of businesses to insure has fallen for two straight years, the underwriters say they are finally pricing losses instead of projections, the Senate is still arguing about whether covering any of it is legal, and in Chicago an insurer is asking a judge to agree that the policy it sold does not answer the claim that arrived.
Total losses have fallen by nearly two thirds in a decade. In the same window one container ship cost the market $2.8bn, the largest marine loss ever recorded. Frequency and severity have come apart, and the two halves of the market disagree about which one to price.
Only 11% of agentic AI use cases reached production last year, and the industry treats that as the problem. Cigna's claim-denial system reached full production and averaged 1.2 seconds of human review. What Zurich, IAG, WTW and Lemonade did differently, and what a court in Minnesota decided it wanted to see.
Insurers spent 2025 and 2026 writing AI out of standard policies and building a new market to replace it. Here is what Armilla, Munich Re, AIUC, Klaimee, HSB and Testudo actually sell, and what it takes to launch a product like theirs.
Book a live, personalized demo with our product team - tell us your use case and see the platform work with your data. No commitment.