He Deleted a Tree on Screen. The Premium Dropped. Now a Billion-Dollar Insurer Says the Old Model Is a Coin Flip
(SeaPRwire) –
By: Logan Pierce
I watched Matt DiStefano delete a tree on screen and watch the premium drop by over $1,100 a year. That’s the hook. The 29-year-old chief science officer at Stand demoed the digital twin tech. He overlaid a heat map on a California neighborhood. Red zones ran along the sides and back of one house, where trees stood too close. He deleted the trees, and the price moved. His argument is that insurers have been pricing wildfire risk wrong. Zip codes give homeowners a score like a credit rating they cannot improve. Stand says it can show you exactly what to fix.
Stand raised $70 million in a Series C round at a $1 billion valuation, led by Saga Ventures. Participants included Lowercarbon, Eclipse, Equal Ventures, Inspired Capital, Neotribe, and Jackson Square Ventures. The company now insures more than $10 billion in property, up from $1 billion just a year ago. CEO Dan Preston claims the business grows about 15 percent a month. That growth has happened in only two states. California handles wildfire. Florida handles hurricanes. Stand sells through brokers and now works with the country’s largest. The small footprint is notable. Two states, two perils, and a billion-dollar valuation.
The digital twin model accounts for house structure, surrounding trees, brush, wind direction, and weather conditions. Stand runs computational fluid dynamics, physics models that DiStefano likened to designing an airplane. He has stints at Boeing and SpaceX on his resume. The company built what it calls the Stand World Model. It runs simulations instantaneously. DiStefano said they successfully taught neural networks to actually understand the physics. The backtest covers roughly 310 California homes. These properties experienced wildfires over three to four years, including the Palisades Fire. The model predicted burn or survival about 85 percent of the time. Competing statistical models got it right 50 to 55 percent of the time.
The criticism of incumbent models is sharp. DiStefano calls them “really a coin flip.” The result is homeowners spending tens of thousands of dollars hardening their houses and still getting non-renewed. He said Stand has heard “hundreds of stories” like that from clients. DiStefano claimed the model has rated some homes at a 99.8 percent chance of total loss. Targeted changes brought that close to 15 percent. The results are Stand’s own. TechCrunch could not independently verify them. That caveat matters. The claim of 85 percent accuracy against 50 percent for incumbents needs external validation. Insurers should not bet real money on it yet.
Preston was previously CEO at Metromile, the pay-per-mile auto insurer, before starting Stand about two and a half years ago. Co-founder Jason Mueller led product at Policygenius. Mueller says his team is about 10 people. They do 100 times the work of his last team of over 100. Modern AI tooling is why. The applied science team is about 10, as well. Mueller described one of Stand’s biggest secrets as making brokers look like heroes. They get near-instant quotes and a platform that replaces trading PDFs. Preston said the company has someone on staff who used to be a construction builder. Mitigation work financing and policy portability are in the works but not short-term.
A billion-dollar valuation on physics-based underwriting is a bet that the next wildfire season will validate the model before the balance sheet catches up, and if it does, every incumbent carrying zip-code risk models will have to rewrite their pricing engines or explain why their customers keep getting dropped.
Author bio: Logan Pierce is an independent business researcher and corporate governance writer on Medium, specializing in insurance tech innovation and venture-stage startup dynamics across North America.