Key insights
- The author argues Reddit is structurally mispriced due to underestimated margins. Q1 2026 EPS beat Q1 2027 consensus, with high incremental net income margins. Catalysts include a potential settlement with Anthropic, setting a precedent for future AI licensing deals, and improving US DAU trends. The author is long RDDT.
Quick update following Q1 and ahead of Q2 earnings: EPS: Q1 2026 EPS came in at USD 1.01. Consensus was USD 0.62. Consensus for Q1 2027 is USD 0.98 - so Reddit just posted a number above what the analysts expect 12 months out, in a quarter with no one-offs. 2026 net income consensus has been revised from USD 871m to USD 1,014m in the past 3 days, but it's still mispriced.
The biggest flaw in consensus: margins. Q1 2026 incremental net income margin was 65.7%, up from 45.5% in Q2 2025. Tax-adjusted floor is around 50% once NOLs run out. Analysts have net income margin stagnating at ~30% - basically at current level - and incremental margin compressing to 35.8% by 2028. Hard to see why - the platform is built, gross margin is 91%, moderators are free labour, Q1 capex was USD 1m. There's no mechanical reason margins compress as revenue scales. Catalysts: the Anthropic ruling found Reddit's claims "qualitatively different from a copyright claim" - fair use isn't a defense to a Terms of Service breach. Settlement plus licensing deal looks probable ahead of Anthropic's IPO, and it sets precedent for the pending Perplexity suit and the Google/OpenAI renewals in H1 2027. S&P 500 inclusion criteria still met. The US DAUq stagnation scare is probably ending - install pop-ups have pushed iOS App Store rank from #162 to #78 in the US over the past 2 weeks (UK #180 to #85, DE #194 to #144).
Worth noting Q2 2026 is the last quarter Reddit reports the logged-in/logged-out DAUq split. Convenient timing to surprise on the metric just before discontinuing it. In any case the split is irrelevant as per my previous posts.
As you can imagine - I'm very long Reddit. Check my previous posts for more detailed DD.