Insights · Insurance AI Weekly

Lemonade’s Q3 in five numbers, and what they say about AI-first distribution

·

,

Lemonade grew revenue 73% last quarter and still lost more money than it did a year earlier. Both facts matter. Together they are the clearest public scorecard we have for direct, app-first insurance distribution built around AI.

What happened

On November 10, Lemonade reported third-quarter 2026 results. Revenue rose 73% to $336.1 million, and in-force premium grew 34% to $1.55 billion, the twelfth straight quarter of growth. Customers rose 23% to about 3.5 million. The net loss widened to about $43.7 million from about $37.5 million a year earlier, and operating expenses rose to about $192 million from $141.2 million. Lemonade still expects its first adjusted EBITDA-positive quarter in Q4 2026. Its car product is now in 14 states, with Florida, Missouri and Pennsylvania added in the last 90 days. (Insurance Journal)

The quarter in five numbers

  • 73% revenue growth. Growth this fast at this size shows that direct digital distribution can scale. It doesn’t need an agent channel to get there.
  • $1.55 billion in force. Twelve straight quarters of growth suggests customers stay and premium compounds. That matters more than any single quarter’s new sales.
  • 3.5 million customers. Customers are growing more slowly than premium. That points to more premium per customer, through bundling and moving into car insurance.
  • $192 million in operating expenses. Spending is still rising fast. Buying customers directly is expensive, and growth spending can hide efficiency gains in the core operation.
  • Q4 adjusted EBITDA target. This is the test. If Lemonade gets there, it is the first real evidence that the model can grow and turn profitable.

Our take

Lemonade’s story is often told as an AI story, but the numbers read more like a distribution story. Direct-to-consumer insurance swaps agent commissions for marketing spend. AI keeps the cost of serving each customer low, but it does not make customers cheaper to win.

For carriers, MGAs and agencies, there are two lessons. First, AI tends to improve the expense ratio before it shows up in the loss ratio. Expect operational savings early and underwriting results later. Second, digital distribution gets cheaper as each customer buys more products. A business that can only sell one policy per customer struggles to make the economics work.

What to watch in Q4

  • Adjusted EBITDA. Whether it actually turns positive, and how much of that comes from cost discipline rather than growth.
  • Car insurance loss ratios. New states usually start with higher losses before pricing data matures.
  • Premium per customer. Continued growth here would support the bundling thesis.

Whatever Q4 shows, Lemonade is still the most transparent test case for anyone weighing a direct, AI-heavy distribution model against traditional channels.

← All Insurance AI Weekly posts

Get the weekly brief

Get our short weekly roundup of insurance software and AI news in your inbox.

Putting AI to work in insurance?

We help insurers, MGAs and agencies modernize core systems, get their data ready and adopt AI responsibly.