Bear (PaulWBryant) | April 1, 2026 Quarter ending: Dec-2025 | Reported: Feb 25, 2026 Market cap: ~$0.96B | P/S: 0.61x | Revenue: $397.0M (+21.5% YoY)
Root delivered record revenue (397M), recordPIF(481, 869), andreturnedtonetprofitability(5.3M) after Q3's warrant-driven loss. Management kept every single Q3 promise — all eight — which earns real credibility. FY2025 was a genuine milestone: $1.5B revenue, $40M record net income, and $192M free cash flow for a company that nearly went bankrupt three years ago.
But the numbers beneath the surface are telling a more complicated story. Revenue growth decelerated for the fifth consecutive quarter (21.5%, down from 67.7% five quarters ago). All three GAAP margins compressed YoY. The gross accident-period loss ratio inflected upward to 62.8% — the worst reading in a year. Premium per policy declined YoY for the first time in available history. And management proactively guided for lower net income in 2026 as they enter a deliberate reinvestment cycle.
This is not a broken company — far from it. But it's a company where the theory ("AI-driven pricing creates a structural compounding advantage") hasn't yet matched the recent numbers (decelerating growth, compressing margins, rising loss ratios). I could be wrong — and the valuation at 0.6x P/S means the market is pricing in a lot of the risk already — but I need to see the 2026 investment phase actually produce reacceleration before I can build conviction.
Thesis: Watchlist — Interesting but unproven at this stage of the cycle.
This is my first formal ROOT analysis. I'm constructing reasonable expectations based on Q3 FY25 data, management's Q3 promises, and the trajectory.
| Metric | Prior Belief (pre-Q4) | Actual Q4 FY25 | Verdict |
|---|---|---|---|
| Revenue | $395-405M (+22-25% YoY) | $397.0M (+21.5% YoY) | In line — low end of range; growth decel continues |
| QoQ revenue growth | +2-4% (improvement vs Q3's +1.3%) | +2.4% ($9.2M) | In line — marginal QoQ acceleration, nothing dramatic |
| PIF | 475-490K (management promised acceleration) | 481,869 (+3.3% QoQ, +16.2% YoY) | In line — healthy, >2x Q4 FY24 pace as promised |
| Net income | $10-20M (seasonal headwind but no warrant catch-up) | $5.3M | Below — profitability thinner than expected |
| Operating income | $15-25M | $10.5M | Below — growth reinvestment hit harder than assumed |
| Combined ratio (net) | 97-101% | 99.7% | In line — at worse end of range, barely sub-100% |
| Accident-period loss ratio | 60-63% | 62.8% | In line but trending wrong — upper bound, worst in 4 quarters |
| Premium per policy | $1,550-1,600 | $1,531 | Below — first YoY decline (-3.3%); mix shift larger than expected |
| Adj EBITDA | $35-45M | $28.8M | Below — margin compression steeper than expected |
| IA channel growth | Continued strong | Tripled YoY new writings | Beat — remarkable acceleration |
| Management credibility | High (Q3 promises trackable) | All 8 promises kept | Beat — perfect scorecard |
Key delta: Profitability came in materially below what I'd have expected. Revenue was fine. PIF was fine. But the operating/net income miss and EBITDA compression tell me the investment spend is running hotter than the revenue lift — at least for now. The premium-per-policy decline is a new concern; management frames it as structural advantage, but it mechanically means each customer generates less revenue, and the loss ratio is not declining to offset it.
| Q1_FY24 | Q2_FY24 | Q3_FY24 | Q4_FY24 | Q1_FY25 | Q2_FY25 | Q3_FY25 | Q4_FY25 | |
|---|---|---|---|---|---|---|---|---|
| Cal date | Mar-24 | Jun-24 | Sep-24 | Dec-24 | Mar-25 | Jun-25 | Sep-25 | Dec-25 |
| Revenue ($M) | 254.9 | 289.2 | 305.7 | 326.7 | 349.4 | 382.9 | 387.8 | 397.0 |
| YoY % | 263.6% | 286.6% | 165.1% | 67.7% | 37.1% | 32.4% | 26.9% | 21.5% |
| QoQ % | 30.9% | 13.5% | 5.7% | 6.9% | 6.9% | 9.6% | 1.3% | 2.4% |
| Gross margin (GAAP) | 34.7% | 34.2% | 39.6% | 41.3% | 41.2% | 39.1% | 38.3% | 38.1% |
| Op margin (GAAP) | 2.1% | 1.3% | 11.3% | 10.7% | 6.8% | 7.1% | 0.1% | 2.6% |
| Net margin (GAAP) | -2.4% | -2.7% | 7.5% | 6.8% | 5.3% | 5.7% | -1.4% | 1.3% |
| Adj EBITDA ($M) | 15.1 | 12.1 | 41.6 | 43.1 | 31.9 | 37.6 | 33.7 | 28.8 |
| EBITDA margin | 5.9% | 4.2% | 13.6% | 13.2% | 9.1% | 9.8% | 8.7% | 7.2% |
| Net income ($M) | (6.2) | (7.8) | 22.8 | 22.1 | 18.4 | 22.0 | (5.4) | 5.3 |
| PIF | — | — | — | 414,862 | 453,800 | 455,493 | 466,320 | 481,869 |
| Prem/policy ($) | — | — | — | 1,584 | 1,614 | 1,616 | 1,581 | 1,531 |
| Gross loss ratio | — | — | — | 56.8% | 56.1% | 58.0% | 58.5% | 59.3% |
| APLR | — | — | — | 61.4% | 55.5% | 59.3% | 59.5% | 62.8% |
| Gross combined ratio | — | — | — | 90.6% | 94.0% | 94.3% | 101.3% | 98.8% |
| Net combined ratio | — | — | — | 91.5% | 95.6% | 95.2% | 102.1% | 99.7% |
| FCF ($M) | 14.3 | — | — | 64.8 | 26.8 | — | — | 65.9 |
| SBC ($M) | 4.6 | 3.8 | 4.3 | 5.8 | 6.4 | 8.4 | 11.6 | 13.7 |
The accident-period loss ratio (APLR) is the single most important metric for an insurance company because it measures the actual pricing accuracy on current-period business. Here's what I see:
| Quarter | APLR | Direction |
|---|---|---|
| Q4_FY23 | 63.7% | — |
| Q4_FY24 | 61.4% | Improving |
| Q1_FY25 | 55.5% | Best in history |
| Q2_FY25 | 59.3% | Deteriorating |
| Q3_FY25 | 59.5% | Deteriorating |
| Q4_FY25 | 62.8% | Deteriorating |
The pattern is clear: APLR bottomed in Q1 FY25 and has deteriorated for three consecutive quarters. It's now back near the Q4 FY23 level (63.7%) — which predates most of the AI pricing improvements management touts. Management guides for 60-65% long-term and Q4 is within that range, but the trajectory is wrong.
Let me be clear: If the AI pricing models improved LTV by 20% in the last 12 months as claimed, I'd expect the APLR to be improving, not deteriorating. There are three possible explanations:
I don't know which is dominant. Q1 FY26 should help disambiguate — if APLR drops back below 60% with seasonal tailwinds, then explanations 1+2 hold. If it doesn't, explanation 3 gains weight.
Net combined ratio of 99.7% means Root is making less than a penny of underwriting profit per dollar of premium. Compare to Q4 FY24's 91.5% — an 8.2 percentage point YoY deterioration. The gross combined ratio (98.8%) is slightly better, suggesting reinsurance drag.
For context, Progressive runs combined ratios in the low-90s consistently. Root's thesis is that better pricing technology should eventually produce consistently superior combined ratios. We haven't seen that yet — FY24 was the only year with sub-95% combined ratios, and FY25 already regressed.
| Quarter | Prem/Policy | Direction |
|---|---|---|
| Q4_FY24 | $1,584 | — |
| Q1_FY25 | $1,614 | Rising |
| Q2_FY25 | $1,616 | Rising |
| Q3_FY25 | $1,581 | Declining |
| Q4_FY25 | $1,531 | Declining (-3.3% YoY) |
First YoY decline in available history. Management frames this as a feature: "as pricing models get better at segmenting risk, can lower premiums for the right customers." In other words, better algorithms let them charge less to good drivers.
I have two concerns with this framing:
GPW declined 6.6% QoQ ($387.2M to $361.6M) even as PIF rose 3.3%. Written premiums lead earned premiums by roughly one quarter. This suggests Q1 FY26 earned premium growth could moderate, which would pressure revenue.
Promise tracker: 8 for 8. Every Q3 promise was kept. I want to give credit where it's due — this is a management team that says what it's going to do and does it.
| Q3 Promise | Q4 Resolution |
|---|---|
| Q4 PIF acceleration | Kept — PIF growth >2x Q4 FY24 pace |
| ~$5M incremental marketing | Kept — S&M +$8.4M YoY |
| ~5 ppts seasonal loss ratio headwind | Kept — modest headwind visible |
| IA growth / no slowdown | Kept — IA tripled YoY new writings |
| Broadly rate adequate | Kept — APLR within 60-65% target |
| Tariff impact not in data | Kept — no claims inflation from tariffs |
| Partnership increasing % of book | Kept — partnerships nearly half of Q4 new writings |
| October PIF accelerated | Kept — consistent with Q4 print |
That's genuinely impressive. CEO Timm navigated from near-death in 2022 to here — that's a turnaround story few can match.
However: The 2026 guidance concerns me. "Lower net income" and "long-term mindset, prioritizing durable value over short-term reporting results" is the kind of language that can mean anything from "we're investing wisely" to "margins are going to erode for longer than you think." I'd feel better if management gave even a rough quantitative framework for what the investment phase looks like.
New Q4 FY25 promises to track:
| Commitment | Specificity | Testable By |
|---|---|---|
| Accelerating annual PIF growth 2026 | Directional | FY26 |
| Q1 2026 elevated shopping, sequential PIF growth | Directional | Q1 FY26 |
| Q1 2026 loss ratio improvement vs Q4 | Directional | Q1 FY26 |
| 2026 loss ratio within 60-65% | Range | FY26 |
| Lower full-year net income 2026 | Directional (down) | FY26 |
| All contiguous states by end of 2027 | Specific | FY27 |
| Toyota partnership: crawl/walk/run trajectory | Qualitative | Multi-year |
Management introduced a formal five-lever framework this quarter. Let me assess each:
1. Pricing (better math = lower prices = market share): In theory, excellent. In practice, the evidence is mixed. 20% LTV improvement claimed but APLR is deteriorating, not improving. Premium per policy declining. The "structural advantage" argument requires patience and faith that the loss ratio catches up. Shows what I know — maybe it will. But the numbers don't confirm it yet.
2. Geographic expansion (36 states to all contiguous by 2027): Reasonable. 80% U.S. population covered already. Incremental states have smaller populations, so marginal revenue impact per state addition diminishes. The regulatory approval process is slow but Root has demonstrated capability.
3. Independent agents (~10% penetrated, tripled YoY): This is the most compelling near-term lever. $100B+ TAM, only 10% penetrated, 3x YoY growth, and the Goosehead API integration demonstrates genuine product-market fit. If there's a reason to be bullish on ROOT, it's this channel.
4. Connected ecosystem (Toyota, Hyundai): Exciting but explicitly early-stage. Management's own analogy — "crawl/walk/run similar to Carvana's 4-year journey" — tells me this is a 2028+ revenue story, not 2026. The Hyundai "Secure+" trademark filing (per scuttlebutt) is a concentration risk worth watching.
5. Direct distribution machine: Incrementally improving (three consecutive quarters of direct new writings growth) but this is the most competitive channel. Root competes here against Progressive, GEICO, and State Farm who collectively spend billions on brand advertising. Root's ML-driven marketing is smart but subscale.
My ranking: IA channel > Pricing > Geographic > Connected ecosystem > Direct. The IA channel is where the numbers actually match the theory right now.
From the pre-computed scuttlebutt:
Customer experience remains a liability. NAIC complaint index ~2x industry average. Android app at 3.2/5. Claims handling complaints persistent. This matters because retention drives the economics — if Root churns 40%+ of customers annually (typical for auto insurance), the embedded acquisition cost advantage is partially offset. Management doesn't discuss retention metrics publicly, which is itself a tell.
Employee sentiment poor. Glassdoor 2.8/5, 39% recommend. Below industry average. Not terminal, but not evidence of a great culture either.
Competitive landscape tightening. Incumbents have all deployed telematics programs. Root's edge is the embedded distribution channel and ML pricing quality, not telematics as a concept. This is worth acknowledging — the moat isn't where the bull narrative says it is.
Analyst consensus tepid. 2 Buy, 4 Hold. Price targets 89−165 with median ~$120. No strong conviction from sell-side.
| Metric | Value | Assessment |
|---|---|---|
| Market cap | ~$960M | Micro-cap |
| P/S (TTM) | 0.61x | Very cheap for any growth company |
| P/E (run-rate Q4) | 45.2x | Misleading — based on thin $5.3M quarter |
| PEG | 0.028x | Absurdly low on paper |
| EV/Adj EBITDA (run-rate) | ~8.3x | Reasonable for a growing insurer |
| Rule of 40 | 38.1 | Below 40 threshold |
| Cash | $669M | 70% of market cap |
The valuation is objectively cheap. $669M in cash plus $391M in investments against a $960M market cap means the insurance operating business is being valued at essentially nothing (or slightly negative when you subtract the $200M debt and $126M preferred redemption value).
The bull math (not mine, but worth stating): If Root can grow to 2B + earnedpremiumata93140M underwriting profit. Add $25M investment income on float. At 15x earnings = $2.5B market cap. The upside is real if the model works.
The bear math (closer to my concern): Combined ratios hover around 99%, net income stays thin or negative through the investment cycle, growth continues to decelerate as the embedded channels saturate initial partners, and Root ends up as a $1-1.5B revenue insurer growing mid-teens with minimal profitability. At 0.5-0.7x P/S that's roughly where we are today.
My take: The valuation provides a floor. Cash alone is nearly 70% of market cap. You're not going to lose your shirt buying ROOT at 0.6x P/S unless the business fundamentally deteriorates. But "cheap with a floor" is not the same thing as "compelling buy." I don't invest based on hope that the investment cycle works — I want to see evidence.
Atlas had ROOT at Conviction 2/5 after Q3 FY25, with the verdict that "Q4 reacceleration data is the make-or-break test."
Q4 results: PIF reaccelerated as promised. Revenue growth continued to decelerate (21.5% vs 26.9%). Combined ratios improved QoQ (99.7% vs 102.1%) but deteriorated significantly YoY. Net income positive but thin.
Where I agree with Atlas:
Where I add nuance:
Watchlist. No position.
The turnaround story is real. The management team has earned credibility. The valuation provides downside protection. The IA channel is genuinely exciting.
But the numbers have to match the theory, and right now they don't — not fully. Growth is decelerating, margins are compressing, the loss ratio is moving the wrong direction, and management is telling me 2026 will be a step back on profitability. I'm not going to pay for that with my capital while the thesis remains unproven at this stage.
I could be wrong. If the IA channel continues to triple and the APLR drops back below 58% in Q1-Q2 FY26, I'll revisit with more conviction. The option value here is real — I just want to see it start converting to intrinsic value.
Bear