AppLovin runs an AI-driven performance-advertising engine for mobile apps and, increasingly, for e-commerce. Its core product is AXON — a self-learning ad engine (built on the old AppDiscovery/MAX auction stack plus the Adjust measurement business) that matches an advertiser's budget to the right user impression in real time, then optimises for return on ad spend automatically. What sets it apart is the economics: roughly 900 employees turn about US$6.2bn of trailing revenue into ~78% operating margins and ~$3.2bn of free cash flow, because the software does the work that armies of ad-ops staff do elsewhere. AppLovin sold its legacy games/apps studio in 2025 to become a pure advertising-software business, and is now opening AXON to e-commerce advertisers via a self-serve dashboard — a market management pegs at 5–10× the size of mobile gaming. For a reader: think of it as a machine that rents out an AI ad-buyer, priced today like one of the market's premium software compounders.
Lifecycle: High-Growth software (Communication Services / ad-tech). Trailing revenue grew ~59% year-on-year to ~US$6.2bn, with ~78% operating margins and ~64% net margins — the profile of a scaled, high-growth platform, not an early-stage one. Post the 2025 divestiture of its games studio, AppLovin is a pure advertising-software business, so we score it on software metrics: revenue growth, gross/FCF margin, Rule of 40, ROIC and moat — not P/E-of-a-mature-company or asset-based lenses.
| Sub-signal | Value | Read | Score |
|---|---|---|---|
| Revenue trajectory | +59% YoY (~$6.2bn TTM) | Elite for a $130bn-cap; Q1-26 rev $1.84bn, sequential re-acceleration | 92 |
| Gross margin | ~88% | Software-grade; cost of revenue is thin infra/rev-share | 90 |
| Operating / net margin | ~78% / ~64% | Extreme operating leverage — ~900 employees run a $6bn+ platform | 95 |
| FCF generation | ~$3.2bn FCF, ~52% FCF margin | FCF/OCF ~0.99 — near-total cash conversion, minimal capex | 93 |
| Balance sheet | Net debt ~$1.1bn; cover ~23× | Term-loan funded buybacks; healthy, not a risk | 72 |
| Earnings quality (§7b) | Clean ≈ reported | Non-op income NEGATIVE — NOT inflated by mark-to-market gains | 88 |
Moat score ≈ 68/100. The flywheel and cost edge are strong; switching costs and pricing power are the softer walls — an advertiser unhappy with attribution can move budget to Meta or The Trade Desk overnight, which is exactly the vector the Competitive Environment below (and the short thesis) targets.
| Rival | Where it competes | Share trajectory vs APP | Threat |
|---|---|---|---|
| Meta (Advantage+ / Audience Network) | The dominant performance-ad AI; e-commerce & app budgets | Incumbent; APP taking share at the margin but Meta is 10×+ the ad revenue | High — also a data source APP is accused of scraping |
| Alphabet / Google (Ads, AdMob) | Search + display + app; measurement rails | Incumbent; APP a challenger in-app | High — platform + policy risk (TOS) |
| The Trade Desk (TTD) | Open-internet DSP / CTV; the “independent” comp | TTD stumbled in 2026 (down ~44% YTD); APP the relative winner in ad-tech | Medium — different lane (CTV/open web) but overlaps on brand/e-comm budget |
| Unity (Grow/ironSource) | Mobile-game monetisation & UA | Weakened post-merger; APP clearly ahead | Low |
| Short-seller thesis (structural) | The business model itself | n/a | Latent — alleges ROAS overstatement, retargeting “attribution theft”, and persistent identity graphs breaching Meta/Google TOS. Unproven; company retained Quinn Emanuel for an independent review; business kept compounding and it entered the S&P 500. Carried as a Bear leg, not a fired gate. |
ROIC is very high (asset-light, ~54× fixed-asset turnover; reported ROE is distorted by buyback-shrunk equity and is not the right lens). Capital allocation has been aggressive buybacks funded partly by term debt — accretive while the stock compounds, but it is why book value/tangible equity is thin (P/B ~55× is a red herring for a software name and should be ignored). Management skin-in-the-game is meaningful (founder-CEO Adam Foroughi); SBC is a genuine cost to watch but not at a gate-tripping level. Quality pillar: 84/100, confidence 72% — the haircut vs a 90 is entirely the unresolved legitimacy/durability question around the attribution engine, not the financials.
The whole call sits here. AppLovin is an elite business at a rich price. On our disciplined, rate-and-growth-warranted anchor it is squarely in the Expensive band, which fires the Valuation-Ceiling gate and caps every horizon at HOLD. The mitigants — a cheap-looking forward multiple, a fat analyst target, and a multiple that is near the LOW end of APP's own 2-year range — are real, but they order the name within the Expensive band; they cannot lift it out of it. That is the anti-hype design working as intended.
| Input | Value | Basis |
|---|---|---|
| Risk-free (10-Y UST) | 4.67% | FRED DGS10, 29 Jul 2026 |
| Equity risk premium | 4.50% | fixed global constant |
| Risk add-on | +2.0% | beta 2.48 > 1.6 → top add-on (dominates r) |
| Discount rate r | ~11.2% | 4.67 + 4.50 + 2.0 |
| g (yrs 1–5) | 15% | secular-growth cap; consensus haircut 25% — disciplined, not the 59% headline |
| g (terminal) | 3% | long-run nominal GDP |
| Warranted P/E | ~20.4× | two-stage; capped at the 26× Comm-Svcs guardrail (raw < cap) |
| Actual clean multiple | ~32.6× trailing | clean EPS ≈ reported (§7b); reported P/E 33.9× |
| Actual ÷ Warranted | 1.60× → EXPENSIVE | ≥1.40× → <40 band |
Implied-growth read (narrative colour, not the score): at $390 on ~11.2% r, the market is implying roughly ~28% durable 5-yr growth; our disciplined estimate is ~15% (a proven >20% grower could justify up to 20%). So the price embeds more growth than we will underwrite — the gap IS the Expensive read.
| Relative cross-check (orders within the band) | Read | Signal |
|---|---|---|
| Forward P/E ~18× (EPS fwd $21.67) | Cheap-looking — but only IF the ~90% next-yr EPS jump lands; that is the bet, not a fact | Attractive-if-delivered |
| PEG ~1.1 (fwd) | Fair for the growth | Fair |
| Own 5-yr history | Decile ~2 — near the LOW end of its own range (traded 30–53× fwd in H1-26); the stock is down 48% from $745 | Cheap vs itself |
| Sector / peers | TTD fwd ~12–30×; APP fwd ~18× mid-pack for ad-tech | Fair |
| FCF yield (FCF/EV) | ~2.3% ($3.2bn / $136.8bn EV) | Expensive (1–3% band) |
| Analyst targets (30 cover) | Mean $654.6 / median $660 / high $860 / low $406 — price ~40% below mean → large “upside” | Bullish (but targets FALLING: last-mo $575 vs last-yr $693) |
| Grades consensus | 23 Buy / 2 Hold / 1 Sell; FMP health B (ROE/ROA 5, but P/E 2, D/E 1) | Bullish herd |
Valuation pillar: 38/100 (top of the Expensive band), confidence 70%. Not lower, because the forward multiple, the down-48% own-history position and the analyst support are genuine cushions; not higher, because the warranted anchor AND the industry guardrail both say Expensive, and the Gate independently caps the signal.
The dominant external force tethered to APP is the AI-performance-advertising cycle — the same AXON engine that drove the 10× run, now extending from mobile gaming into e-commerce. It is a genuine secular tailwind, but three things cap it to Neutral (no amplification): mobile/DTC ad-spend is cyclical and macro is a headwind; the attribution/privacy/TOS overhang is unresolved; and APP is a high-beta member of the very AI cohort the macro report's concentration tail is armed against.
| Horizon | Read | Source / date |
|---|---|---|
| Historical | Explosive — AXON took AppLovin from a games publisher to a $130bn ad platform | 2023–25 results |
| Current | Decelerating tape: ad-tech de-rated hard in 2026 (TTD −44% YTD, APP off 48% from high); armed concentration tail; attribution overhang live | 30-Jul MacroDriver; price action |
| Forward | E-commerce self-serve (Jun 2026) + gen-AI creative = large TAM, but unproven at scale and platform-dependent | Company, Jun–Jul 2026 |
Driver score 56/100 → Neutral. In the 36–64 band, so it does not amplify: it leaves the base HOLD unchanged (it could neither lift a BUY to STRONG BUY nor push a SELL to STRONG SELL). Thesis-invalidation floor: Q2 evidence that ROAS is deteriorating or the e-comm ramp is stalling, or a Meta/Google data-access restriction — any of which would turn this driver into a live headwind.
Regime is Stagflation-lite (energy shock re-armed on Iran/Hormuz re-escalation, Brent ~$90; policy-tight into cooling growth; narrow, contested, tape-unconfirmed leadership). AppLovin maps to Communication Services (XLC), which the 30-Jul macro rates Underperform (short) / Underperform (medium) / Neutral (long) — mega-cap ad/concentration weakness. Net pressure is a Headwind on the short and medium horizons (the amplification-relevant anchor), easing to Neutral long. Because a long entry here would be fighting that headwind it is framed Contrarian, low conviction (40). Crucially, the armed “S&P 500 concentration / AI earnings-quality unwind” tail materially applies to APP as a high-beta AI-ad mega-mover — it does not fire DNB (armed, not triggering) but it is why the Bear carries a cohort de-rating leg. Pressure did NOT enable any amplification (base is HOLD).
Source: sector-map (XLC) — no dedicated APP watchlist signal · Macro report 2026-07-30
The tape is against a buyer today. Confluence across timeframes is strongly bearish. Price ($390) is below the daily SMA20 ($452), SMA50 ($492) and SMA200 ($521); the weekly and daily trends are down/strong-down; the stock just broke recent support and sits only ~9% above its 52-week low of $359. RSI (daily 37.5, weekly 41) is soft but not yet washed-out oversold (<30) and MACD histograms are negative — so this reads as a falling-knife/continuation, not a confirmed capitulation bottom.
| Signal | Reading | Score |
|---|---|---|
| MTF trend confluence | Monthly up (fading, MACD rolling) but Weekly/Daily/Hourly/15m down → ~32 (All-Bearish edge) | 32 |
| Risk-reward / position-risk | Below all key MAs; nearest logical stop ($359) is close, but no higher-low reversal to lean on yet | 35 |
| Relative strength | Underperforming SPY and XLC on 1m/3m (down 48% from high) | 28 |
| Macro overlay (low sensitivity) | Fed on-hold/tight, VIX elevated (energy shock), XLC out of favour | 30 |
| Sentiment | Analyst grades bullish (23 Buy) but targets being CUT and price action bearish — mixed | 50 |
| Catalyst | Q2 earnings 5 Aug (binary, <7 days) — focused but high path-risk | 45 |
Timing pillar: 34/100 (Weak), confidence capped at 40% by the Earnings-Event gate. A buyer has no technical edge here — the reachable entries are a confirmed higher-low bounce off the $359–377 zone, or a post-earnings reclaim of the $450s on volume; chasing into the print is the worst of both.
| Date | Event | Impact | Forecast | Previous | Relevant? | Why |
|---|---|---|---|---|---|---|
| 2026-08-05 | AppLovin Q2-2026 earnings | High | EPS $3.72–3.76e; rev $1.90–1.95bn (co. guide) | Q1 EPS $3.56 | ✅ Yes | The single biggest near-term catalyst; first read on the e-commerce self-serve ramp |
| 2026-08-01 | US Nonfarm Payrolls | High | — | — | ⚠️ Indirect | Sets the risk-on/off tone + rate path that drives high-beta growth multiples |
| 2026-08-13 | US CPI | High | — | — | ⚠️ Indirect | Inflation surprise moves the 10-Y that anchors APP's warranted multiple |
| Date | Event | Actual | Forecast | Surprise | Impact |
|---|---|---|---|---|---|
| 2026-07-29 | Concentration-tail breadth tell | RSP flat vs XLK +5.5% | — | Narrow | Bearish for high-beta AI cohort |
| 2026-07-29 | Iran/Hormuz re-escalation | Brent ~$90 (+8%) | — | Risk-off | Energy-shock re-armed → Stagflation-lite |
Two macro prints (payrolls 1 Aug, CPI 13 Aug) bracket a company-specific binary (earnings 5 Aug) — a genuinely cluttered 2-week window. High path-risk both directions; the disciplined stance is to let the print resolve rather than pre-position into it.
| Timeframe | Trend | Direction | RSI | MACD | Key S/R | Breakout | Vol |
|---|---|---|---|---|---|---|---|
| Monthly | Uptrend (fading) | ↗︎ | 50.1 | +, hist rolling over | S 359 / R 745 | Res breakout (stale) | 0.96× |
| Weekly | Downtrend | ↘︎ | 41.3 | −, falling | S 320–364 / R 489–576 | Support breakdown | 1.18× |
| Daily | Strong downtrend | ↘︎ | 37.5 | −, hist −3.3 | S 387–418 / R 515–622 | Support breakdown | 1.32× |
| Hourly | Strong downtrend | ↘︎ | 38.2 | − | S 378–395 / R 409–421 | Support breakdown | 0.67× |
| 15-min | Strong downtrend | ↘︎ | 42.9 | − turning | S 378–387 / R 404–409 | — | 0.12× |
| Confluence: Strongly Bearish · MTF Score 32 | |||||||
Only the stale monthly candle is still nominally an uptrend, and its MACD is rolling over. Weekly through 15-min are all down with support breakdowns. Price is pinned below every key moving average and ~9% above the $359 52-week low. There is no higher-timeframe support to lean on until $359–364; a bounce needs a confirmed higher low or a post-earnings volume reclaim of the $450s to become tradeable. Until then, rallies are sells and the path of least resistance is down.
AppLovin daily close, 19 May → 30 Jul 2026. The late-May peak near $613 has unwound to ~$390 — a 48% drawdown from the $745 52-week high — with price now below all key moving averages and testing the $359–390 support shelf into earnings.
Q2 (5 Aug) beats and management quantifies a fast e-commerce self-serve ramp; ROAS durability rebuts the short thesis; the 10-Y eases and market breadth broadens (RSP catches SPY) so the concentration tail disarms. The multiple re-rates back toward ~24–28× forward as the ~90% EPS growth is validated. ~+33% to ~$520. This is the ‘everything works and the macro cooperates’ path — real, but it needs both the company AND the tape.
Q2 is solid-but-not-euphoric; the e-comm ramp is early and guidance is measured. The stock stays range-bound roughly $360–450 as elite fundamentals fight a rich starting multiple, an XLC headwind and high beta. Forward EPS grows into the multiple rather than the multiple expanding. Net roughly flat, ~$405. A HOLD is the honest read: you own a great business already fairly-to-richly paid for.
Two compounding legs. (1) Cohort de-rating (systemic): the armed ‘S&P 500 concentration / AI earnings-quality unwind’ tail triggers — an AI markdown / hyperscaler capex guide-down / breadth break (RSP<SPY) — and high-beta APP's forward multiple compresses from ~18× toward ~10–12× (trailing ~33× → ~18–20×), a 40–50% multiple move independent of the company. (2) Idiosyncratic: Q2 shows ROAS/attribution softening OR Meta/Google restrict data access OR the short-seller PIG allegations gain regulatory/platform traction — cutting forward EPS estimates on top of the compression. Together they break the $359 52-week low decisively toward ~$245 (−37%). Falsification: breadth broadens (RSP catches SPY) and Q2 confirms ROAS durability — either largely defuses this leg.
Forecast: 0 of 3 groups met → Wait. No entry edge today: the name is above our fair-value anchor, the tape is a strong downtrend, and a binary print lands in 5 days. The two reachable paths are (a) a Catalyst entry — a post-5-Aug move >+5% with raised guidance on >2× volume; or (b) a Technical entry — a confirmed higher-low off $359–377 OR a volume reclaim of the $450s. Fundamental would only open on a pullback toward the $245–300s or a materially higher fair-value estimate post-earnings.
Forecast: No exit trigger is live (this is a non-held HOLD). For an existing holder the operative discipline is the $359 stop and the Q2 thesis check on 5 Aug.
Position sizing not computed — specify your portfolio allocation and role for sizing guidance.
{
"ticker": "APP",
"date": "2026-07-31",
"version": "v6",
"brand": "",
"company": "AppLovin Corporation",
"currency": "USD",
"exchange": "NASDAQ",
"exchange_ticker": "NASDAQ:APP",
"isin": "US03831W1080",
"api_ticker": "APP",
"finder_ticker": "APP",
"finder_exchange": "NASDAQ",
"analysis_status": "starting",
"price_at_rating": 390.21,
"signal_short": "HOLD",
"signal_medium": "HOLD",
"signal_long": "HOLD",
"primary_signal": "HOLD",
"short_hold_reason": "expensive",
"short_entry_confirmed": false,
"lifecycle_stage": "high_growth",
"quality_score": 84,
"quality_detail": {
"industry_benchmark_name": "Rule of 40 (software)",
"industry_benchmark_value": 111,
"industry_benchmark_score": 96,
"moat_score": 68,
"roic_percentile_vs_peers": 90,
"capital_allocation": 70,
"management_skin_in_game": 68,
"earnings_quality": "clean\u2248reported; non-op income negative"
},
"valuation_score": 38,
"valuation_detail": {
"fcf_yield": 2.3,
"implied_growth_rate": 28.0,
"consensus_growth_rate": 35.0,
"historical_valuation_decile": 2,
"forward_pe": 18.0,
"trailing_pe_clean": 32.6,
"warranted_multiple": 20.4,
"actual_multiple": 32.6,
"val_multiple_basis": "clean trailing P/E",
"discount_rate_r": 0.112,
"risk_free_10y": 0.0467,
"g_near": 0.15,
"g_term": 0.03,
"warranted_ratio": 1.6,
"val_band": "expensive",
"sector_guardrail_pe": 26,
"guardrail_breach": true
},
"timing_score": 34,
"timing_detail": {
"mtf_confluence": 32,
"risk_reward_score": 35,
"relative_strength_vs_spy": -18.0,
"relative_strength_vs_sector": -12.0,
"catalyst_clustering_score": 45,
"dynamic_macro_weight": 0.1
},
"driver_score": 56,
"driver_label": "Neutral \u2014 no amplification",
"econ_stance": "Contrarian",
"econ_pressure": "Headwind",
"econ_conviction": 40,
"xlc_signal": "U/U/N",
"overall_confidence": 40,
"fair_value_est": 245,
"stop_loss": 359,
"target_price": 405,
"scenario_base_target": 405,
"scenario_bull_target": 520,
"scenario_bear_target": 245,
"scenario_weights": {
"bull": 20,
"base": 55,
"bear": 25
},
"entry_groups_met": 0,
"entry_conviction": "Wait",
"exit_groups_live": 0,
"exit_action": "Hold",
"gates_triggered": [
"Gate 3 Valuation Ceiling (Expensive: 32.6\u00d7 \u2265 26\u00d7 guardrail AND 1.60\u00d7 warranted)",
"Gate 2 Earnings Event (Q2 5 Aug; timing conf capped 40%)"
],
"do_not_buy_triggers": [],
"dnb_armed_not_firing": [
"Trigger 2b: S&P 500 concentration / AI earnings-quality unwind tail is ARMED (not triggering) \u2014 caps at HOLD via Valuation Ceiling, carried as Bear cohort-de-rating leg"
],
"next_update_date": "2026-08-06",
"next_update_basis": "earnings 2026-08-05 +1d"
}
First-ever Donatien report on AppLovin. Signal HOLD / HOLD / HOLD — an elite business (Quality 84, Rule of 40 ~111, clean earnings) capped by the Valuation-Ceiling gate: at ~32.6× trailing clean P/E it is above both the 26× Comm-Services guardrail and 1.60× its warranted multiple. No Do-Not-Buy fires — the armed AI-concentration tail is a Bear watch, not a live catalyst. Entry conviction Wait (0/3). Re-underwrite after Q2 earnings on 5 Aug.