Nirvana and Why Trucking Insurance Is Becoming a Data Pricing Problem

Every renewal season, a small carrier opens the envelope from its insurer and finds the number went up again. Not because the fleet crashed more. Often it didn't crash at all. ATRI's operational cost data shows liability insurance premiums hit a record 10.2 cents per mile in 2024, roughly 10 percent of total operating cost, after rising 37.8 percent over the past decade. Over that same period, heavy-duty truck injury and fatal crash rates declined. Premiums and safety have decoupled, and that decoupling is the clearest evidence available that trucking insurance is not pricing risk. It is pricing uncertainty.

The burden lands unevenly. ATRI found that small fleets pay two to three times more per mile for liability coverage than large fleets, not because they crash two to three times more often, but because they generate too few miles for an underwriter to model confidently. Where the data is thin, the insurer prices in the uncertainty. A five-truck operation running clean for a decade is, in the underwriter's spreadsheet, still mostly a guess, and the guess is priced against the fleet.

Why the pool keeps getting more expensive

Two forces drive the whole structure upward. The first is litigation. ATRI's forensic analysis of trucking verdicts found the median nuclear verdict, defined as an award above $10 million, reached $36 million in 2022, roughly 50 percent higher than a decade earlier. The broader trend is accelerating: 135 nuclear verdicts landed against corporations in 2024, a 52 percent jump in a single year, totaling $31.3 billion. In more than 80 percent of verdicts above $1 million, non-medical damages ran up to ten times the plaintiff's actual medical bills. Against this, the federal minimum primary liability for general freight is $750,000. A single median verdict burns through primary coverage and most fleets' entire excess tower. The market has responded rationally and brutally: the cost of the $10 million to $15 million excess layer rose roughly 45 percent in just four years.

The second force is that the insurers themselves are losing money. Commercial auto has posted an underwriting loss in 15 of the past 16 years. In 2024 the line's combined ratio was 107.2, meaning insurers paid out $1.07 in claims and expenses for every $1.00 of premium collected, a $4.9 billion underwriting loss in one year. This number matters more than any other in understanding the opportunity. An incumbent losing seven cents on every premium dollar cannot cut prices to defend market share. Its only lever is raising rates on the pool, which is exactly what pushes its best customers toward anyone who can price them individually.

What Nirvana actually does

Nirvana, founded in 2021 by Rushil Goel after he built the fleet business at Samsara, underwrites commercial trucking using the data the industry already produces. Its models are trained on more than 30 billion miles of real-world telematics, combined with FMCSA regulatory records, capturing speed, braking, cornering, route selection, and time-of-day exposure at the level of the individual fleet. Safe operators can earn upfront discounts of up to 20 percent for connecting their telematics, and the company says its underwriters reach decisions up to 15 times faster than traditional carriers.

The data layer compounds on the claims side. Nirvana reports adjusting claims 30 to 40 percent faster than incumbents and has paid some physical damage claims in hours rather than weeks. For a small carrier, that speed is not service quality, it is utilization: a truck waiting on an adjuster is a fixed-cost asset producing zero revenue. And when liability is contested, telemetry and video give fleets a defense they never had. In a legal environment where substance-related negligence findings increase expected awards by over 340 percent and venue alone can move a median verdict from $2.5 million in federal court to $3.6 million in state court, evidence has become survival infrastructure.

The selection flywheel

Here is the mechanism that makes this a structural story rather than a product story. Usage-based, data-verified pricing selectively attracts the safest fleets, because they gain the most from being measured accurately. Every safe fleet that leaves a legacy pool makes that pool statistically worse, pushing legacy rates higher, which pushes the next tier of fleets to connect their telematics and get quoted. Adverse selection, the force that normally kills insurers, becomes the growth engine of the one holding the better model. In a line running a 107 combined ratio, the insurer that can select even modestly better than the pool average is not competing on margin. It is competing against counterparties who lose money by default.

The capital markets are pricing that flywheel in. Nirvana crossed $100 million in premiums under management in 2024, doubling year over year, with what it describes as best-in-class loss ratios backed by top-tier reinsurers. It raised an $80 million Series C at roughly $830 million in March 2025, then a preemptive $100 million Series D at $1.5 billion nine months later. A valuation that nearly doubles in under a year, on rounds the company says it did not need, is investors betting that selection advantages in insurance compound rather than erode.

What could break it

The risks deserve the same numerical honesty. First, the cycle: commercial auto's losses came during a period of rising rates, and an AI-native book that has only grown through a hard market has not proven its loss ratios hold as the book ages and pricing softens. Second, reach: the model requires telematics access, and the riskiest fleets are the least likely to connect, which caps the addressable share of a market where lawsuits are still growing 3.5 percent annually. Third, the moat: telematics data is abundant and incumbents run usage-based programs of their own. Nirvana's edge is architectural, an underwriting, claims, and safety stack built around the data from day one rather than bolted onto a legacy core. That kind of retrofit is something incumbents announce far more often than they finish, but fifteen consecutive years of losses is also fifteen years of motivation.

The larger repricing

Read alongside what Netradyne is doing on the safety data side, the structure is now visible end to end: driver behavior becomes structured data, structured data becomes an insurance rate, and the rate becomes a per-mile cost advantage or a slow-motion bankruptcy. At 10.2 cents per mile, insurance is a top-three operating cost with no spot market and no hedge. The only instrument a fleet has is proof, and proof is now a data feed. The carriers that can produce it will be priced on their own record. The ones that cannot will keep paying for everyone else's.

Sources: ATRI Operational Costs of Trucking (2025) and litigation forensic analysis, AM Best, Marathon Strategies, TechCrunch, FreightWaves, Crunchbase News, SiliconANGLE, company disclosures.