Wall Street had Reddit ($RDDT) in the “show me” pile heading into Q1. After Friday’s print, that pile got a lot smaller.
The numbers were not subtle:
- Revenue: $663 million vs. $611 million expected (+8.5% beat)
- Year-over-year revenue growth: 69% — from $392M a year ago
- EPS: $1.01 vs. $0.58 expected (+74% beat)
- Net income: $204 million, up from $26 million — a 680% jump
- Q2 guidance: $715M–$725M revenue (above the $712M consensus)
The stock responded the way you’d expect: +13.82% on Friday, the biggest single-day move since the post-IPO honeymoon.
What changed
Two things, both bigger than they look. One: AI data licensing. Reddit’s deal with Google to license its archive for training has rolled into a recurring revenue line that’s now meaningful — and Reddit is one of a tiny handful of platforms that owns 20 years of high-quality, human-generated, conversational text. That’s a moat no LLM can replicate.
Two: ad density without user revolt. Reddit’s average revenue per user is still roughly a quarter of what Meta extracts from a Facebook user, but the ad load on the platform has crept up dramatically over the last 18 months without the typical “this place is ruined” backlash. That’s a unique cocktail of community tolerance and ad-tech execution.
The reaction on Wall Street
Multiple firms raised price targets after the print. The common thread in the upgrade notes: analysts had been modeling Q1 revenue around $610M and Q2 guidance around $700M. The actual print blew through both, and the implied 2026 revenue trajectory is now closer to $3.0B than the previous $2.6B consensus. That’s a $400M upward revision, which on a still-growing platform translates to a meaningful step-change in 2027 earnings models.
The Play
The first-derivative play — just buying $RDDT — isn’t bad, but the stock has already ripped 13.8% and option premiums are ugly. The second-derivative trade is more interesting: which other companies own proprietary user data that LLMs need to train on?
The candidates aren’t obvious. $POET (POET Technologies) and other niche-data platforms have been buying social/behavioral datasets and licensing them to enterprise AI buyers. $SOFI sits on financial behavior data that any consumer-facing AI agent eventually needs. Even X (privately held but expected to file in late 2026) is suddenly a more interesting comp.
The macro thesis: every LLM company is short data, and there are maybe a dozen public platforms that own the kind of data they need. Reddit just told you what one of those platforms is worth on the new multiple. The rest of the cohort hasn’t priced it in yet.