Lemonade launched Lemonade Car in Florida on 26 August and in Missouri on 2 September — two states in eight days. Its Florida release frames the move the way every state launch gets framed: with that licence, Lemonade Car is "now available in states representing nearly 50% of the U.S. car insurance market."
That is a market-size claim, and market size is the least sensitive assumption we carry on the line. Take the addressable fleet in our Lemonade model from the 8 million vehicles it holds to one billion — more vehicles than exist anywhere — and base fair value moves from $53.55 to $53.93. Thirty-eight cents is not an artefact of the number we picked. It is the ceiling on what that assumption can ever be worth.
Move the other lever on the same line — the ceiling on how many Car policies Lemonade adds per quarter — from 38,000 to 50,000 and fair value goes to $55.72. To 70,000 and it goes to $59.28. The two levers are not in the same weight class, and a state launch is an argument about the wrong one.
What the numbers are, and whose they are
- Car is the largest single swing factor in the model. Remove the line entirely and base fair value falls from $53.55 to $40.79 — $12.77, or 23.8% of the whole model. The vertical matters enormously. Its market-size assumption does not.
- The 8-million-vehicle fleet is ours, defined as private passenger vehicles in the states Lemonade is licensed in or filing for, not the national fleet. It is an assumption, not a disclosure.
- The policy count is ours too, and it is the softest figure here. Lemonade publishes in-force premium for Car — $239 million at 30 June, as reported on 29 July — but not a Car policy count. Ours, 114,409 policies and 10,877 net adds in the June quarter, is derived from that premium at about $2,089 a policy. Every "policies per quarter" figure below inherits that derivation.
- The 38,000-a-quarter add ceiling is ours as well, and it is roughly three and a half times the pace the line is running at now.
- Lemonade discloses none of this by state. There is no per-state premium, policy count or loss ratio, so nothing here is a claim about Florida's book specifically.
Market size is a brake, not an engine
The reason the fleet assumption is inert is structural rather than a matter of calibration.
Our Car line grows the way a subscription grows: a quarterly rate of net new policies that ramps toward a ceiling, damped by the share of the market already taken. Market size enters that arithmetic in exactly one place — the damping term. At 114,409 policies against 8 million vehicles, Lemonade holds 1.43% of its own addressable fleet, so the market-size assumption is withholding 1.43% of each quarter's adds and doing nothing else. Make the fleet infinite and you recover that 1.43%. That is the entire lever, and it is why the ceiling is 38 cents.
Nor does the brake tighten later. Run the published assumptions out over the model's twenty-quarter horizon and the line ends at about 528,000 policies against a fleet that has itself grown — 6.2% penetration. Saturation never comes near binding, so where the wall sits never matters.
That is the general shape of the thing. For a company holding one or two percent of the market it addresses, market size is a brake, not an engine. It can only ever subtract, and at 1.4% penetration it is barely touching the pedal.
The two curves
Move each assumption by the same proportion and the curves separate on the first step.
Triple the fleet and our fair value gains 25 cents. Triple the quarterly add ceiling and it gains $13.36 — fifty-three times as much for the same proportional move. One line is pinned to the axis with a hard ceiling 38 cents above it. The other is a straight climb with no ceiling at all.
The right-hand end of the add-ceiling line is not a forecast, and we are not offering it as one: three times a 38,000 ceiling would be roughly ten times the pace the line runs at today. It is drawn to show the shape of the curve. The finding is the shape, not the endpoint.
The cleanest way to state it is as an equivalence. Everything the market-size assumption could ever be worth — the full 38 cents, at any fleet size you like — is also what you get by lifting the quarterly add ceiling from 38,000 to 40,100. That is 2,100 policies a quarter, less than a fifth of what the line added in the June quarter on our own derivation.
What can actually raise the rate
Two things from the last fortnight bear on the rate, and neither of them is a map.
The first is cross-sell. On the second-quarter call, co-founder Shai Wininger said Car sales to existing Lemonade customers were the highest ever, and that cross-sell customers "typically represent between 40% to 50% of new to Lemonade car sales." Roughly half of Car's new business is being sold to people who are already Lemonade customers in states the company already operates in. That half needs no new licence at all.
The second is Missouri. On 2 September Lemonade launched Car there alongside Lemonade Autonomous Car, which gives Tesla drivers 50% off every mile driven using Full Self-Driving (Supervised). Wininger called the simultaneous launch "a first for us." Our Car line already carries an autonomous product as one of its two streams and values it at zero, because until this month it was not sold anywhere — and it stays at zero here, on one state with no disclosed policy count and no disclosed premium. What it is, structurally, is a price product rather than a footprint product: the kind of thing that changes how many people buy, not how many people could.
Florida is the bigger market and Missouri got the newer product. On our model's mechanics, only one of those two launches is operating on a lever that pays.
What we are not claiming
Two disclosures, because both cut against the piece.
Our model was struck on 24 August, two days before the Florida launch, against a $50.71 price. Its fleet assumption already covers states Lemonade is "licensed in or filing for," and Lemonade had plainly been filing in Florida. Florida may already be inside the 8 million. That is precisely why the figures above are a bound rather than an update: the 38-cent ceiling holds whether Florida is counted already or not.
And this is the point where the analysis usually goes hunting for Florida's registered private-passenger fleet. Florida publishes 17,226,196 autos and pickups against 24,368,726 registrations of every class — trailers, vessels, motorcycles, buses and mobile homes included — for registrations current in May 2025. It does not split private-passenger from commercial use, and pickups sit in with the cars. We are not going to manufacture that split, because the answer does not depend on it. Set our fleet to Florida's 17.2 million and the model gains 20 cents. Add those 17.2 million on top of the 8 million already there and it gains 26. Make it a billion and it gains 38. Every number that could go in that box lands between 20 and 38 cents, so the precision nobody has is precision nobody needs.
So the model does not change, and the size of the "no" is the finding. What would change it is the third quarter showing Car adding policies at a materially faster clip — a number Lemonade does not print, which is why we derive it, and why the derivation is the first thing to argue with.
What to watch
- Car in-force premium in the third quarter. It was $239 million at 30 June, 16.7% of Lemonade's $1,434 million total. The sequential step is the only observable that speaks to the add rate, and it is the number that would move our ceiling rather than our fleet.
- Whether growth spend converts. Management said growth spend rises sequentially in the third quarter behind Car and seasonal renters strength, defended at roughly 3x lifetime value to acquisition cost. Spend is the input to the rate; if premium added per dollar of spend falls, the lever that pays is getting more expensive.
- Florida's loss experience, not Florida's size. Florida is the hardest personal-auto market in the country, and Car is where Lemonade's underwriting has improved most — the blended gross loss ratio was 60% in the June quarter with a record-low 5% loss-adjustment expense ratio, Car's own at 7%. A Florida book that pushes the Car loss ratio back up costs more than the entire market-size lever is worth.
- A second state for Autonomous Car. It exists in one. A second and a third would make it a rate product rather than a pilot, and would be the first thing on this line worth re-underwriting.
Lemonade's Florida launch on 26 August, including the "nearly 50% of the U.S. car insurance market" framing, and its Missouri launch and Autonomous Car description on 2 September, are from the company's own releases. June-quarter in-force premium, loss ratios, guidance, growth-spend commentary and the cross-sell quote are as Lemonade reported on 29 July and said on its second-quarter call, captured in our June-quarter notes. Florida's registration counts are the state's Current Registered Vehicles report for registrations current in May 2025; the state does not separate private-passenger from commercial use. Every fair value, sensitivity, ceiling and equivalence here is R40 arithmetic on our own Lemonade model, struck 24 August 2026 against a $50.71 price — including the 8-million-vehicle fleet, the 38,000-policy quarterly add ceiling, and the Car policy count and net adds, which Lemonade does not publish and which we derive from the in-force premium it does. Model assumptions are our estimates, not company guidance.
