Date: 2026-02-22 Quarter: Q3 FY25 (ended Sep-25) Analyst: Saul Rosenthal
Thesis: Tentatively Intact — Watching
74% revenue growth at this stage is exceptional by any measure. But I have to be honest with you — this company has only been public for 7 quarters, the FY26 margin guide is a dramatic step-down, and NRR is declining. Those are three things I do not take lightly. The numbers this quarter were really, really good. The forward guide raises legitimate questions. I'm interested but not yet convinced. Let me take you through it.
No prior position in ARX. First formal review. Coming in fresh, no anchoring.
| Q1 FY24 | Q2 FY24 | Q3 FY24 | Q4 FY24 | Q1 FY25 | Q2 FY25 | Q3 FY25 | |
|---|---|---|---|---|---|---|---|
| Fiscal | Q1 FY24 | Q2 FY24 | Q3 FY24 | Q4 FY24 | Q1 FY25 | Q2 FY25 | Q3 FY25 |
| Calendar | Mar-24 | Jun-24 | Sep-24 | Dec-24 | Mar-25 | Jun-25 | Sep-25 |
| Revenue ($M) | 128.0 | 142.0 | 153.5 | 161.9 | 181.7 | 239.0 | 267.4 |
| YoY % | — | — | — | — | 41.9% | 68.3% | 74.2% |
| QoQ % | — | +10.9% | +8.1% | +5.5% | +12.2% | +31.5% | +11.9% |
| Adj EBITDA ($M) | 12.8 | 15.5 | 25.9 | 25.1 | 27.0 | 46.6 | 105.0* |
| EBITDA Margin | 10.0% | 10.9% | 16.9% | 15.5% | 14.9% | 19.5% | 39.3%* |
| EWP ($M) | ~580 | ~680 | ~795 | ~885 | ~890 | ~955 | 1,042.9 |
| Members | ~180 | ~195 | ~204 | ~224 | ~238 | ~252 | 265 |
| NRR | — | — | 151% | — | 146% | 151% | 135% |
| Gross Loss Ratio | — | — | 54.7% | — | — | — | 50.1% |
*Q3 EBITDA includes 39Minvestmentgain.Normalized: 66M, ~24.7% margin.
Revenue grew 74% year-over-year. Let me say that again. 74% year-over-year. From $153.5M to $267.4M. On an accelerating trajectory — 42% → 68% → 74%. This is not a fluke. The trajectory has been relentlessly upward for three consecutive quarters of comparable data.
Exchange Written Premium crossed $1 billion for the first time — $1.04B, up 31% YoY. Members grew 30% to 265. The Gross Loss Ratio improved from 54.7% to 50.1% — meaning the underwriting is getting better as the platform scales, not worse. That is the flywheel working.
The Q3 EBITDA print looks spectacular at $105M (39% margin), but I want to be precise: that includes a $39M investment gain. Strip it out and you get roughly $66M in operating EBITDA — still a 24.7% margin, still excellent, still accelerating year-over-year. I follow the money, the results.
NRR came in at 135%. That is strong — I'd take 135% in any business. But I'm watching it closely because it was 151% two quarters ago. Declining NRR while the company is still in hyper-growth could mean many things. On the call, management attributed it to the platform mix shift — more third-party insurers joining relative to owned/mission members — which mechanically changes how NRR is calculated. If that's the explanation, and the business is still expanding rapidly, I can accept it. But I'll be watching this number like a hawk.
This is where I need to be careful and not get starry-eyed about the Q3 numbers.
Management guided FY26 to 5.0B + EWPat5.4269M EBITDA. Let me put that in context:
Meanwhile, revenue will grow — they're guiding EWP to nearly 5x the Q3 quarterly run rate. But the margin is deliberately being compressed from ~25% now to 5.4% in FY26. That is a massive margin step-down.
The strategic logic is: they're shifting from writing risk on their own balance sheet (high margin, capital-intensive) to becoming a pure exchange for third-party insurers (lower margin on revenue, but asset-light, scalable). Third-party insurer revenue was 27% → 32% → 39–40% target. This is actually a smart platform play — lower margins, but lower risk, less capital, and potentially much larger TAM.
I get the logic. I've seen it before. The question is: Do I trust management to execute this transition without stumbling? Seven quarters of public history does not give me enough data. The execution risk is real.
I don't usually pay too much attention to data moat claims. Every company says they have one. But let me tell you what ARX has built:
The network effects here are real. A bad actor who inflates claims poisons the risk pool for everyone — so members have an incentive to underwrite well. This is structurally different from a typical insurance company. I like it.
Clean. $2.55B in cash and investments, $122M in debt. The $2.5B in investments isn't idle cash — it's insurance float supporting the risk they're retaining. Still, the fortress balance sheet means zero financial distress risk and they have the capital to fund the third-party transition without dilution. That matters.
Atlas scored this 3.5/5 — I think that's fair. The cheap valuation (2.5x EV/TTM revenue at 74% growth), the 66% drawdown from highs creating an entry opportunity, and the unique platform model are all real. Atlas correctly flagged the FY26 margin compression as the central risk and the 7-quarter history as a conviction limiter.
Where I'd add nuance: the NRR decline deserves more weight. 151% to 135% in two quarters is not trivial. And the investment gain distorted Q3 optics in a way that could mislead casual observers.
Prior: No view. Coming in fresh.
Updated: This is a genuinely interesting business that I would not have looked at before (insurance!), but the platform economics are more like a software marketplace than a traditional insurer. The 74% revenue growth with an accelerating multi-quarter trajectory is exactly what I look for. The balance sheet is immaculate. The data moat appears real — 57k risk attributes per member and a network with disciplined membership is not easily replicated.
But I am not ready to own this yet. Only 7 quarters of public history. The FY26 margin guide is a dramatic step-down and I need to see management execute through that transition without missing. NRR declining needs to stabilize. The exchange take rate (8%) is meaningful but thin — any pricing pressure would hit EBITDA hard.
Action: Watch. Do not buy yet.
If Q4 FY25 and Q1 FY26 confirm: (a) NRR stabilizes ≥130%, (b) GLR continues below 52%, (c) new insurer pipeline converts — I would revisit with much higher conviction.
At 3.8x P/S and 74% growth, the market is giving this away relative to what it would be worth if FY26 guidance is delivered. The risk/reward is interesting. But I've learned over the years that "interesting" is not the same as "ready to own." I need another quarter or two.
Best, Saul