Filed: 2026-02-22 | Period: Q4 FY25 (Oct–Dec 2025) | Persona: muji (CMF_muji)
Links: Press Release | Scout Brief | Atlas Analysis
AXON is the most capital-efficient platform in software — $1.7B/quarter revenue, 84% EBITDA margin, 79% FCF margin — and e-commerce is the next TAM unlock that could re-accelerate growth back to Tier 1.
| Q1 FY24 | Q2 FY24 | Q3 FY24 | Q4 FY24 | Q1 FY25 | Q2 FY25 | Q3 FY25 | Q4 FY25 | |
|---|---|---|---|---|---|---|---|---|
| Mar-24 | Jun-24 | Sep-24 | Dec-24 | Mar-25 | Jun-25 | Sep-25 | Dec-25 | |
| Revenue ($M) | 678 | 711 | 835 | 1,152 | 1,158 | 1,245 | 1,423 | 1,658 |
| YoY % | — | — | — | — | +71% | +75% | +70% | +44% |
| QoQ % | — | +5% | +17% | +38% | +1% | +7% | +14% | +17% |
| EBITDA ($M) | — | — | — | 722 | — | — | — | 1,391 |
| EBITDA Margin | — | — | — | 63% | — | — | — | 84% !! |
| FCF ($M) | — | — | — | — | — | — | — | 1,312 |
| FCF Margin | — | — | — | — | — | — | — | 79% !! |
FY25 Full Year: Revenue $5,484M [Non-GAAP] +70% YoY | EBITDA $4,513M [Non-GAAP] 82% margin | FCF $3,949M 72% margin
Q1 FY26 Guidance: 1, 745–1,775M revenue [Non-GAAP] (~52% YoY ^^) | EBITDA 1, 465–1,495M (~84% margin)
Here's the thing about AppLovin that most people miss: AXON isn't an ad network — it's a prediction engine trained on 1.6 billion daily active users and 1 million+ advertisers.
That's the crowdsourced intelligence flywheel. Every ad auction, every conversion, every LTV signal from every advertiser gets fed back into a single unified model. The more advertisers use it, the better it gets at predicting outcomes for every OTHER advertiser. This is compounding data moat, not just software.
Building block, not process tool. AXON is embedded in advertiser bidding infrastructure. It's not a dashboard they log into — it's the engine executing their bids in real-time. Switching cost is real ROI destruction.
Self-serve GTM. No SI dependency. Advertisers onboard, connect their data sources, and the algorithm does the rest. This is why EBITDA margins hit 84% — there's no professional services drag on the model.
Cloud-native, software-only post-divestiture. When they sold the apps business, they stripped out all the hardware and content cost. What's left is pure algorithm + infrastructure. The margin trajectory from 58% to 84% EBITDA over 8 quarters tells the whole story.
Revenue $1,658M, +44% YoY. The YoY deceleration is real — down from +75% in Q2 FY25 — but the sequential trajectory is actually fine: +1% → +7% → +14% → +17% QoQ, accelerating within FY25.
The YoY comp is getting harder. Q4 FY24 was $1,152M after the AXON 2.0 inflection. Now they're lapping that inflection. This is a comp problem, not a business problem.
EBITDA margin 84% !! — extraordinary. The operating leverage in this model is unlike anything in software. Every incremental dollar of revenue drops to EBITDA at ~84 cents. FCF margin 79% means the cash is real — not optical.
Rule of 40: 44 + 79 FCF% = 123 !!!. Atlas said 150 using EBITDA — either way, off the charts.
The original AXON business was mobile gaming. ~70% of revenue historically from mobile game publishers bidding for installs. The e-commerce expansion is the attempt to replicate that playbook for the $1T+ e-commerce advertising market.
Q4 update:
NUANCE: 57% go-live vs 30-day payback is the right metric to watch. If e-commerce advertisers break even in 30 days, the spending flywheel self-funds. They don't need to be convinced — the math does the convincing. This is how AXON scaled in mobile gaming and it's exactly the pattern replicating in e-commerce.
The Q1 FY26 guidance re-acceleration (+52% YoY vs +44% in Q4) is partly explained by easier comps but also suggests e-commerce contribution is ramping. Watch for e-commerce revenue disclosure in H1 2026 — that's when we'll know if the TAM expansion is real.
AXON already had the best bidding algorithm. The remaining friction: advertisers still had to produce their own ad creative. If your ad creative is mediocre, even the best bidding algorithm underperforms.
Gen AI creative tools remove that friction:
This is the same "convenience over price" thesis that drives enterprise adoption. When APP can say "we generate your creatives AND optimize your bids AND you break even in 30 days," the sales cycle collapses. DING DING DING.
The call included pointed questions about Meta (DSP overlap) and CloudX (Google's new offering). Management pushed back hard on both.
On Meta: Meta is a publisher AND a platform. They bid for impressions on behalf of advertisers but they're also competing with those advertisers for eyeball attention. APP is pure platform — pure middleman, zero publisher conflict. Different alignment structure.
On Google CloudX: New entrant. Has reach but not the advertiser-specific model depth. The AXON moat is the years of per-advertiser LTV data. A new entrant can't replicate that overnight.
The real competitive question is: can anyone else aggregate 1.6B DAUs + 1M advertisers into a unified model? Not Meta (publisher conflict), not Google (different architecture), not The Trade Desk (no owned inventory). APP's moat is structural, not just technical.
The capital allocation is excellent. They're not hoarding cash or doing bolt-on acquisitions — they're returning capital while maintaining balance sheet flexibility. At $4.5B EBITDA run rate, net debt is a non-issue.
| Metric | Value | Note |
|---|---|---|
| DAU | 1.6B | Training the AXON model [Non-GAAP/Operational] |
| SBC % Revenue | 5% FY25 | Down from 21% FY23 !! |
| Net Debt / EBITDA | 0.23x | Immaterial |
| Buyback / FCF | 65% | Aggressive capital return |
| 1%→5% Conv Headroom | +400bps | Organic TAM from conversion optimization |
| E-com Go-Live Rate | 57% | Product working |
| E-com LTV Payback | 30 days | Self-funding flywheel |
Prior: No prior muji coverage of APP. Starting fresh.
After this review:
1. The YoY deceleration narrative is misleading. Sequential momentum is actually re-accelerating: +1% → +7% → +14% → +17% QoQ within FY25. The YoY story is purely about lapping the AXON 2.0 inflection from FY24. The business is NOT slowing down internally.
2. E-commerce is not priced in. The current P/S of 23.7x is pricing in mobile gaming maturation. If e-commerce scales to even 20-30% of revenue by FY27, the TAM and growth rate look completely different. The optionality is not in the multiple.
3. 1%→5% conversion headroom is underappreciated. Management cited this repeatedly — the ability to optimize conversion rates from current ~1% toward 5% is incremental revenue without adding a single new advertiser. Pure algorithm improvement → revenue growth.
4. SBC discipline is real. 21% of revenue in FY23 → 5% in FY25. This is management prioritizing shareholder value. Most growth companies go the other way.
Current Tier: 2 (moving toward 1)
Hold / Add on weakness. wsm is already long APP 6.8% + 6.4% LEAPS (Jan'28 $450C) — appropriate sizing.
APP is one of the most capital-efficient software businesses in existence. 84% EBITDA margin with +44% growth and a credible e-commerce expansion thesis is an extraordinary combination. The deceleration is comp-driven, not structural. Q1 guide re-accelerates.
The LEAPS at $450C for Jan'28 are the right expression — they capture the e-commerce optionality without paying full premium today. If e-commerce GA in H1 2026 shows meaningful revenue contribution, I'd consider adding to the LEAPS position.
Key things to watch Q1 FY26:
Risk: If e-commerce go-live rate stalls or advertisers see LTV payback extend past 30 days, the re-acceleration thesis breaks. That's the sell signal.
long APP 6.8% + 6.4% LEAPS
-muji
Atlas: Conviction 4/5. Six-factor: Growth Strong, Trajectory Decelerating, Margins Elite, Dominance Dominant, Valuation Fair (EV/Rev/Growth 0.38x — cheaper than Meta growth-adjusted), Special Present.
Muji vs Atlas:
Analysis saved: 2026-02-22