Back to Blog
UPDATE

Q4 2025 earnings are in. Reddit beat every metric — revenue $726M (+70%), EPS $1.24 (+32% beat), DAUq 121.4M (+19%) — and the stock rose 4.5% after-hours, though Q1 guidance deceleration signals forward concerns.

Read our Q4 2025 earnings update
4-Lens AnalysisRDDTTechnology

Reddit: 74% Revenue Growth, 55% Traffic Decline — What's Really Happening?

Reddit's financials look exceptional on the surface. But Google traffic dropped 55%, data licensing growth collapsed from +66% to +7%, and insiders sold $192.7M with zero purchases. Our four-lens analysis reveals what the headline numbers obscure.

February 4, 2026|10 min read
Revenue Growth
+74%

YoY (Q3 2025)

Traffic Decline
-55%

Google traffic (2022-2025)

Insider Sales
$192.7M

382 sales, 0 purchases

Data Licensing
+7%

Down from +66% (Q1)

Reddit went public in March 2024 with a compelling story: a platform with 19 years of authentic human conversation, uniquely valuable for AI training. The "AI data play" narrative drove significant investor interest — data licensing deals with Google and OpenAI seemed to validate the thesis.

Fast forward to late 2025, and the story has quietly changed. Management now emphasizes that "our core business is ads" — a directional reversal from Q1's claim that Reddit was "the #1 most cited domain for AI across all models."

Meanwhile, the financial performance looks outstanding: 74% revenue growth, positive profitability ahead of expectations, and strong user engagement. So what's the concern?

We ran Reddit through four analytical lenses — Gravy Gauge, Myth Meter, Regulatory Reader, and Insider Investigator — to understand whether the headline metrics tell the full story.

Want the full 4-lens analysis with signal assessments and model debates?

Opus + Sonnet ensemble. 4 lenses. 7 debates. Full evidence citations.

View RDDT Analysis

The Traffic-Revenue Divergence

Revenue
+74%

YoY growth driven by ARPU efficiency

Google Traffic
-55%

Decline from April 2022 to 2025

This is the central puzzle. Reddit's Google traffic — which represents 50-70% of total users — declined 55% from April 2022 to 2025. Yet revenue grew 74% YoY.

How? The answer is ARPU efficiency:

  • ARPU improved 41% YoY through better ad targeting and monetization
  • Price increases offset volume decline
  • Machine learning improvements made each user more valuable

This is excellent execution. But efficiency gains have natural limits. You can't improve ARPU 41% every year indefinitely. At some point, the traffic decline catches up.

The Math Problem
If traffic continues declining at 10-15% annually while ARPU growth slows to 10-15%, you get negative revenue growth. The current divergence cannot persist — the question is when it converges.

What Four Lenses Found

Revenue Durability
CONDITIONAL

Revenue depends on Google traffic Reddit cannot control; 74% growth via ARPU efficiency

Narrative-Reality Gap
DISCONNECTED

Management pivoted from 'AI data essential' to 'core business is ads' within 6 months

Regulatory Exposure
ELEVATED

Active securities litigation + Section 230 dependency + untested data licensing legal basis

Governance Alignment
MISALIGNED

$192.7M sold by top 5 executives, 382 sales, 0 purchases over 6 months

Signal Convergence
Four lenses independently flagged different aspects of the same underlying issue: operational metrics are diverging from the narratives that drove IPO valuation. Insiders monetized $192.7M during this transition period.

The IPO Thesis vs. Reality

Reddit's IPO differentiation thesis centered on AI data value. The pitch: Reddit's corpus of authentic human conversation is uniquely valuable for training AI models. Data licensing deals with Google and OpenAI seemed to validate this.

Here's what happened to data licensing growth:

Q1 2025+66% YoY

"Reddit is the #1 most cited domain for AI across all models"

Q2 2025+24% YoY

Growth slowing but still expanding

Q3 2025+7% YoY

"I'll just remind everybody that our core business is ads"

That's not slower execution — it's narrative reversal. Management went from positioning Reddit as essential to AI to explicitly downplaying the data business within six months.

The remaining data licensing backlog is $173M — approximately 4 quarters at the current run rate. No renewal terms have been disclosed. Investors who bought the "AI data play" thesis now own an advertising company.

What Insiders Did While the Narrative Changed

$192.7M

Total Sold

382

Sale Transactions

0

Purchases

Over six months post-IPO, Reddit's top 5 executives sold $192.7 million in equity through 382 transactions. Not a single executive made a single purchase.

All sales were conducted under Rule 10b5-1 plans — pre-scheduled selling programs designed to avoid accusations of insider trading. These plans are legal and common.

But the pattern is notable: selling accelerated during the period when data licensing growth collapsed from +66% to +7% and management pivoted the narrative. New 10b5-1 plans were adopted in Q3 2025 — during the transition.

Correlation, Not Causation
We cannot establish that insiders possessed material non-public information. The timing correlation between selling, metric deterioration, and narrative pivot is observable. Causation requires evidence we do not have.

The Questions That Determine the Thesis

Can ARPU growth outpace traffic decline indefinitely?

41% ARPU improvement offset 55% traffic decline. But efficiency gains have ceilings. What happens when ARPU growth slows to 10%?

What happens when data licensing backlog depletes?

$173M backlog at current run rate = ~4 quarters. No renewal terms disclosed. Will customers re-up when initial contracts expire?

Is Section 230 dependency existential or manageable?

94% of revenue depends on user engagement enabled by Section 230 immunity. Reform would require fundamental business model changes.

What multiple should an advertising company trade at?

If investors reprice RDDT from "AI data play" to "advertising company," what does fair value look like?

Where Our Models Disagreed

The most substantive debate: Should Revenue Durability be classified as CONDITIONAL or FRAGILE?

Opus Position: CONDITIONAL

Revenue resilience proves management capability. 74% growth despite 55% traffic decline demonstrates execution. Conditions changed and revenue increased — that's the definition of not-fragile.

Sonnet Position: FRAGILE

Traffic decline is a leading indicator of revenue collapse. Historical Gravy Gauge examples show lag between condition change and revenue impact. ARPU gains are masking structural deterioration.

Resolution: CONDITIONAL. The lens definition requires "revenue that disappears when conditions change." Revenue grew 74% when conditions changed. FRAGILE is not demonstrated by current evidence — but the minority position is preserved as a monitoring trigger.

What to Watch

IMMINQ4 2025 Earnings (Feb 5, 2026)

This analysis is based on Q3 data. Results tomorrow may materially update several findings.

HIGHYoY Advertising Revenue Growth

If negative YoY growth occurs, upgrade REVENUE_DURABILITY from CONDITIONAL to FRAGILE.

MEDARPU Growth Rate

Currently +41% YoY. If <10% while traffic decline continues, the divergence closes unfavorably.

MEDTamraz v. Reddit Litigation

Class certification would materially change litigation economics. Dismissal would de-escalate regulatory exposure.

Bottom Line

Reddit's financial execution is excellent; its structural dependencies are concerning. Management turned traffic decline into revenue growth through skillful monetization. That's real value creation.

But the IPO thesis — that Reddit would be essential to AI — has been abandoned by management within six months of going public. Insiders monetized $192.7M during this narrative transition. Multiple correlated risks (Google traffic, Section 230, data licensing legal basis) could be simultaneously stressed in an adverse scenario.

Investor Posture: HIGHER SCRUTINY — Strong operational execution but structurally dependent on factors outside management's control. Q4 2025 earnings tomorrow may materially update several findings.

Full Analysis with Signal Breakdowns

Explore the complete four-lens assessment including debate transcripts, evidence citations, and monitoring triggers.

View RDDT Analysis

This report was generated by the Runchey Research AI Ensemble using primary SEC data and reviewed by Matthew Runchey for accuracy.

This analysis is for educational purposes only and does not constitute investment advice. See our Editorial Integrity & Disclosure Policy and Terms of Service.