Short answer: the sell-off in APP (and arguably PLTR) is less about deteriorating fundamentals and more about a regime shift in how markets price software in the AI era.
But that does not automatically mean “mispriced”. What you are seeing is a transition from growth-multiple valuation to AI survivability valuation. The distinction matters.
Let us unpack this carefully.
---
1. The paradox: exceptional fundamentals, collapsing stock
Your numbers are correct, and the magnitude matters.
AppLovin Q4 2025:
Revenue: $1.658B (+66% YoY)
Net income: $1.102B (+84% YoY)
Adjusted EBITDA: +$1.399B (+82% YoY)
Free cash flow surged with ~84% EBITDA margins
Operationally, this is elite performance for a software platform. The company beat expectations and even issued above-consensus guidance.
Yet shares dropped sharply after earnings (roughly mid-teens to ~20% depending on the session).
This disconnect tells you something critical:
> The market is no longer pricing software primarily on current earnings.
It is pricing future AI positioning risk.
---
2. The “violent technology cycle” Apollo is describing
Apollo’s co-president John Zito is not predicting software collapse. He is describing a valuation reset phase.
Key idea:
AI is forcing a rapid separation between software that becomes infrastructure vs software that becomes replaceable.
Zito characterised the moment as an “extremely violent technological cycle” where:
valuation logic resets,
investors aggressively differentiate winners and losers,
even healthy companies reprice first, fundamentals later.
Private equity data supports this:
AI threatens many niche software business models built on automation or workflow layers.
Deals done at peak SaaS multiples are now under pressure.
Investors fear AI agents can replicate entire software categories.
So markets are asking a new question:
> “Does this company own AI leverage, or is it exposed to AI substitution?”
---
3. Why APP sold off specifically
The sell-off was not random panic. Several overlapping fears appeared simultaneously.
(A) Growth deceleration signal
Even strong companies get punished when growth peaks.
Investors focused on:
expected growth slowing from hyper-expansion levels,
sequential moderation despite strong YoY numbers.
In a normal cycle this is fine.
In a regime shift, it triggers multiple compression.
---
(B) AI competitive uncertainty (the real driver)
Markets are worried that:
AI agents may reduce dependence on traditional ad-tech optimisation layers.
Big platforms (Google AI initiatives, new AI adtech startups) could compress margins.
Even if APP is executing well today, investors fear structural disruption tomorrow.
This is classic early-cycle behaviour:
> Markets price optional future disruption more aggressively than current earnings.
---
(C) Narrative contagion: the “software apocalypse” trade
Recent AI releases triggered a cross-sector repricing:
software stocks broadly sold off,
Nasdaq weakened despite solid macro data,
AI disruption fears spread across industries simultaneously.
This is macro positioning, not company-specific analysis.
---
4. Why this resembles past platform transitions
Historically, during platform shifts:
Cycle Market mistake
Internet (1999–2002) Sold profitable infrastructure with weak dot-coms
Cloud (2015–2016) Mispriced SaaS leaders during transition
Mobile ads (2018 ATT fears) Overestimated ecosystem destruction
Markets initially assume maximum disruption.
Reality later becomes:
some models die,
infrastructure players compound massively.
Apollo’s message aligns with this: software usage will likely increase, but not all software captures value equally.
---
5. Is APP an AI platform or an AI casualty?
This is the core investment question.
Bull case (why it could be mispriced)
APP increasingly behaves like:
an AI optimisation engine,
a data flywheel,
performance advertising infrastructure.
Its AI models are already driving revenue efficiency and margin expansion, according to management.
If AI improves ad targeting rather than replacing it, APP becomes a beneficiary, not a victim.
Markets often misprice companies during narrative shocks. Analysts already called parts of the sell-off “overblown”.
---
Bear case (why markets may be rational)
The risk investors are pricing:
AI agents could eventually:
automate user acquisition,
reduce intermediary optimisation layers,
shift value to platform owners (Google, Apple, model providers).
If optimisation becomes commoditised, margins compress violently.
In other words:
> APP must prove it is an AI owner, not merely an AI user.
---
6. Where PLTR fits into the same debate
PLTR and APP are being grouped together because both sit in the “AI application layer”, not pure model providers.
Markets are asking identical questions:
Question APP PLTR
Proprietary data advantage? Yes (ad performance data) Yes (enterprise + gov data)
AI embedded in workflow? Yes Yes
Replaceable by agents? Unclear Less clear
Platform lock-in? Moderate High
PLTR arguably scores higher on switching costs and mission-critical integration, which is why its valuation resilience has historically been stronger during narrative shocks.
---
7. What is actually happening structurally
We are entering a three-phase AI equity cycle:
Phase 1 (2023–2024)
Everything AI rerates upward.
Phase 2 (now)
Markets panic about AI destroying software. Multiples compress indiscriminately.
Phase 3 (coming)
Capital concentrates into:
data monopolies,
workflow-embedded AI,
infrastructure platforms.
Apollo is essentially saying we are early in Phase 2.
---
8. So: mispricing or correct pricing?
The intellectually honest answer:
Likely cyclical mispricing, not fundamental misunderstanding.
But only if APP proves three things over the next 12–18 months:
1. AI improves margins faster than it commoditises adtech.
2. Revenue growth stabilises despite AI competition.
3. Its models become harder to replicate due to proprietary data scale.
If those hold, today’s sell-off looks like a transition-cycle valuation reset, not a thesis break.
If not, markets are simply early.
---
The key insight
The market is no longer rewarding software for being profitable.
It is rewarding software for being structurally indispensable in an AI-native economy.
APP and PLTR are being stress-tested under that new framework.
Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.
