LMND — Lemonade, Inc.
Lemonade has done in two years what most AI promises never deliver: its gross loss ratio has fallen from 88% to 62% as its data set has scaled, converting a struggling insurtech into a business generating $111 million in quarterly gross profit and posting adjusted free cash flow for three consecutive quarters. At $62 per share — roughly 5.5 times 2026 revenue guidance — the stock prices in the successful completion of a transition that management has scheduled for Q4 2026. Interesting but requires a specific catalyst to be actionable; that catalyst is adjusted EBITDA breakeven by year-end.
The property and casualty insurance industry in early 2026 is experiencing an unusual convergence. After two years of elevated catastrophe losses, inflationary claims costs, and industry-wide underwriting deterioration, the major carriers spent 2024 and 2025 repricing aggressively. Progressive's combined ratio returned to its target range. Allstate came back to profitability after two consecutive years in the red. Rate increases in personal lines outpaced claims inflation for the first time in the cycle. The result is that traditional P&C carriers are generating their best underwriting results in a decade — and the moment has arrived to test whether any AI-first insurer can sustain the structural advantage it claimed to have, or whether the incumbents simply needed a repricing cycle to absorb what they once called a threat.
Against that backdrop, Lemonade's Q4 2025 results are harder to explain away than its critics expected. Revenue grew 53% year-over-year to $228 million. Gross profit grew 73% to $111 million. The gross loss ratio — claims paid as a percentage of earned premium — reached 62%, the best result in company history. Adjusted free cash flow was $37 million positive, the third consecutive positive quarter. The stock had gained 59% over the trailing year entering 2026, then sold off 27% through mid-March on broader market weakness before gapping up on March 17 after a Morgan Stanley upgrade. What follows is an attempt to determine what the business is actually worth and whether the current price is the right price to own it.
The U.S. property and casualty insurance market writes roughly $700 billion in annual premium. It is dominated by State Farm, Progressive, Allstate, GEICO, and a collection of regional carriers whose positions have been largely unchanged for decades. Independent agents and brokers controlled approximately 64% of P&C distribution through 2025 — a figure that has barely moved despite more than a decade of digital-first competition. The economics of insurance are straightforward: collect premiums, invest the float, pay claims, keep the underwriting margin. Structural advantage belongs to whoever prices risk most accurately, distributes most efficiently, and processes claims at lowest cost. For decades, doing all three required large field forces, agent networks, and actuarial databases built over generations. What machine learning applied to underwriting and claims automation potentially does is compress all three cost centers simultaneously — but only at scale, and only if the models improve faster than the underlying risk deteriorates.
Lemonade operates as a full-stack carrier: it underwrites risk directly, retains a portion on its own balance sheet, and cedes the remainder to reinsurers. Through most of its history, it ceded 55% of written premiums through quota-share arrangements — effectively sharing the economics with reinsurers because its loss ratios were not good enough to justify full retention. In July 2025, management reduced that cession to 20%. That decision is the most strategically significant thing Lemonade has done. It is not a financial engineering move — it is management putting its own capital behind the assertion that the improved loss ratios are durable. If they are right, the company pockets roughly $30 million in annualized savings and retains substantially more of its earned premium. If the loss ratios deteriorate, the company absorbs materially more downside from the next major loss event.
The company serves more than 3 million customers across renters, homeowners, pet, car, and life insurance in the United States and a growing European presence. The Giveback program — in which premiums not used for claims are donated to causes customers choose at enrollment — is a behavioral economics mechanism designed to reduce moral hazard by aligning customer incentives with keeping claims low. What it demonstrably does is attract customers who self-select for lower-risk profiles, which is worth something at the underwriting model level. It also drives referral at near-zero acquisition cost in the demographic cohort Lemonade targets most aggressively.
The central claim of the Lemonade thesis is that AI underwriting produces a superior and improving loss ratio as the data set scales. Every policy written, every claim processed, every fraud pattern detected improves the model pricing the next policy. This is the flywheel argument: scale generates data, data improves pricing, better pricing reduces loss ratios, lower loss ratios generate more gross profit, more gross profit funds more growth. The question is whether the flywheel is real — producing a widening competitive advantage — or whether it is being confused with a favorable underwriting cycle that lifted all carriers simultaneously.
The data makes the distinction visible. Lemonade's gross loss ratio improved by approximately 26 percentage points between Q3 2023 and Q4 2025. Over the same period, the broader P&C industry improved its combined ratios by approximately 5 to 8 percentage points through repricing. The differential — roughly 18 to 20 points of outperformance against the industry trend — is difficult to attribute entirely to favorable conditions. Something the model is doing is producing a structurally better result. The comparison to the most operationally efficient traditional carrier sharpens the picture:
| Metric | Lemonade (Q4 2025) | Progressive (2025 est.) | Industry Avg. (2025) |
|---|---|---|---|
| Gross Loss Ratio | 62% | ~67% | ~73% |
| Combined Ratio (est.) | ~105–110% | ~96% | ~98% |
| Gross Profit Margin | 43% | n/a | n/a |
| IFP / Revenue Growth | 31% / 53% | ~15% | ~5–7% |
Lemonade's gross loss ratio of 62% now sits below Progressive's estimated loss ratio — the same Progressive that has spent decades as the acknowledged efficiency standard in personal lines. The expense ratios tell a different story: Lemonade's expense ratio remains elevated because the company is still scaling and still investing in customer acquisition. On a fully combined basis, Lemonade is not yet competitive with Progressive's 96% combined ratio. But the direction is unambiguous, and the trajectory of that combined ratio — driven by the loss ratio falling faster than the expense ratio is rising — is the operating leverage story made visible in the gross margin line.
Full-year 2025 revenue was approximately $738 million. GAAP net loss was approximately $165.5 million, reflecting stock-based compensation and ongoing operating investment. The GAAP figure is real — the dilution from stock compensation is a genuine cost to shareholders — but it obscures the economic trajectory. Adjusted EBITDA loss narrowed to $5 million in Q4 2025 alone, compared to losses measured in tens of millions per quarter two years prior. Net loss as a percentage of gross earned premium has improved from (36%) in Q3 2023 to (14%) in Q3 2025. Adjusted free cash flow was positive for six of the last seven quarters, with $37 million generated in Q4. The reconciling item between GAAP losses and positive adjusted FCF is primarily stock-based compensation, which adds back approximately $25–30 million per quarter in non-cash charges. The adjusted FCF is the better near-term indicator of whether the business is generating real cash; the GAAP loss captures the ongoing dilution shareholders accept as the price of the equity-funded growth phase.
One event in 2025 deserves direct acknowledgment. The January Los Angeles wildfires generated approximately $19.6 million in net losses for Lemonade, plus a $6.9 million California FAIR Plan assessment — roughly $26.5 million in total. At the old 55% quota-share structure, a portion of that loss would have been absorbed by reinsurers. At the new 20% structure, Lemonade retains more of the next major event. This is the risk management accepted in exchange for the better economics of premium retention. A more severe wildfire season, a Gulf Coast hurricane, or a clustering of large events in 2026 would test the profitability timeline and potentially force a re-evaluation of the reinsurance posture. This is not a theoretical risk — it already materialized in Q1 2025 — and it is the most credible scenario under which the current investment thesis fails in 2026.
Daniel Schreiber co-founded the company in 2015 and has served as CEO since January 2024, adding the Chairman role that the board justified on the grounds of unified strategic leadership. He owns approximately 14.7 million shares — roughly 10% of the company — valued at approximately $1.2 billion at current prices. There has been no insider selling reported in the past 18 months. The reinsurance restructuring, the Tesla autonomous car insurance partnership, and the accelerating European expansion are all decisions made by someone with substantial personal economic exposure to the outcome. The Tesla partnership is strategically distinctive: traditional actuarial models have no meaningful historical data on autonomous vehicle accident rates and severity, and an AI-native underwriter that collects real-time vehicle behavior data has a structural head start in underwriting a product class where every incumbent is starting from zero. It is early and small today; the optionality is real.
The growth trajectory is made visible in the following table. These are the variables that, if trending correctly, make the investment case — and if trending wrong, would invalidate it:
| Quarter | IFP ($M) | IFP Growth (YoY) | Gross Loss Ratio | Gross Profit Margin | Adj. FCF ($M) |
|---|---|---|---|---|---|
| Q3 2023 | $719 | 22% | 88% | 19% | ($1) |
| Q4 2023 | $797 | 25% | 85% | 20% | $38 |
| Q1 2024 | $846 | 22% | 83% | 22% | ($6) |
| Q2 2024 | $889 | 24% | 79% | 22% | $16 |
| Q3 2024 | $962 | 27% | 77% | 24% | $5 |
| Q4 2024 | $944 | 26% | 73% | 27% | $48 |
| Q1 2025 | $1,030 | 22% | 73% | 29% | $22 |
| Q2 2025 | $1,083 | 24% | 70% | 30% | $25 |
| Q3 2025 | $1,158 | 30% | 67% | 41% | $18 |
| Q4 2025 | $1,240 | 31% | 62% | 43% | $37 |
Ten consecutive quarters showing two simultaneous trends: IFP growth accelerating from 22% to 31%, and loss ratios improving from 88% to 62%. This is not what the modal skeptical reading of Lemonade expected. The bear case through 2022 and 2023 was that the loss ratio would plateau or worsen as the company moved into more complex, higher-severity product lines such as homeowners and car. It did the opposite. Gross profit grew from $22 million in Q3 2023 to $111 million in Q4 2025 — a five-fold increase over eight quarters while in-force premiums grew 72%. The gap between IFP growth and gross profit growth is the AI operating leverage argument made concrete.
Management guided 2026 IFP growth of approximately 30%, which would put year-end IFP above $1.6 billion. Revenue growth guidance of approximately 60% is driven largely by the reduced reinsurance cession retaining more premium on the income statement. Europe — where IFP grew more than 200% year-over-year in Q2 2025 — represents a growth surface that is earlier in the adoption curve than the United States and carries a different catastrophe profile than the California and Gulf Coast exposures that dominate the domestic book. Lemonade currently serves approximately 3 million customers. The U.S. personal lines market encompasses roughly 140 million households. Effective penetration of the addressable market is approximately 2%. The runway from 2% to 10% penetration, which would put the customer base above 10 million, implies a business several times its current size without requiring geographic expansion beyond existing markets.
At $62 per share — the price at which the stock opened on March 17, 2026 — Lemonade carries a market capitalization of approximately $5.07 billion. Against 2026 revenue guidance of approximately $1.18 billion, that implies a forward price-to-revenue multiple of roughly 5.5 times. The company carries net cash, which reduces the enterprise value modestly. There are no GAAP earnings to compare against, and no normalized pre-tax earnings on which to base a traditional valuation — the business is not yet demonstrating the through-cycle earnings power that would allow a multiple-based assessment of intrinsic value. For a company growing revenue at 60% with gross margins expanding toward 45% and management-guided EBITDA profitability within four quarters, 5.5 times forward revenue is a defensible price. It is not a cheap price. It leaves little margin of safety if the transition slips.
The most credible bear argument is not that Lemonade's AI is illusory — the loss ratio data refutes that — but that the improvement is partially cyclical and will not hold through a severe catastrophe year, particularly now that the reinsurance buffer has been substantially reduced. The 2023–2025 improvement coincided with a P&C repricing cycle that benefited all carriers; Lemonade's outperformance of the industry trend by 18 to 20 percentage points is compelling but not proof that the advantage is impervious to a bad year. A single severe Atlantic hurricane season or a repeated California wildfire event in 2026 could spike the retained loss ratio, delay the EBITDA timeline by one to two years, and reprice the stock substantially. The answer is that this risk is inherent to any insurance carrier and is not unique to Lemonade — the question is whether the underlying underwriting advantage is real enough to justify owning the exposure. The data through Q4 2025 says yes; 2026 claims experience will be the first full test at the new reinsurance structure.
The catalyst is adjusted EBITDA breakeven in Q4 2026. If the gross loss ratio holds in the low-to-mid 60s through a normal claims year and the $5 million adjusted EBITDA loss from Q4 2025 converts to a positive number by Q4 2026 as guided, the stock has a clear path to re-rating as a profitable growth compounder rather than a speculative insurer. An investor with high confidence in that outcome and a 12-to-18-month horizon would find the current price reasonable. An investor who requires proof before commitment should watch the Q2 and Q3 2026 results: if the loss ratio holds and adjusted EBITDA continues its narrowing, the thesis is on track and the current price will look attractive in hindsight.
The flywheel is real. The question is whether the investor needs to own it before the evidence is complete — and at 5.5 times forward revenue, the market is asking them to decide.
Was this analysis useful?
Related Companies