Larry Ellison nailed it:
The real AI moat isn’t the model.
It’s the proprietary data.
Claude, ChatGPT, Gemini, Grok… all trained on mostly the same public internet data.
→ Outputs converge. Models get commoditized.The winners? Companies sitting on unique, real-world datasets competitors can’t copy.•
$HIMS — Millions of personalized health interactions + MedMatch AI feedback loops
• $DUOL — Billions of language learning exercises driving unmatched personalization
• $LMND — Real-time behavioral + claims data powering AI Jim & Maya
• $META — Massive owned social & engagement data across billions of usersThe software apocalypse narrative is wrong.
Data-rich verticals will keep pulling ahead, not because their foundation models are better, but because their data is irreplaceable.
Next wave of AI value = private & proprietary. What’s your top data-moat stock?
Daniel@danielisdizzy
Larry Ellison just exposed a fundamental truth about AI: Models like Claude, ChatGPT, Gemini and Grok are all trained on the same public internet data. When the input is the same, the output converges. The moat isn’t the model. It’s the data. Companies training on proprietary datasets have an advantage competitors cannot replicate. That’s why the “software apocalypse” narrative is wrong. And why $HIMS, $DUOL, $LMND, $META — companies that own unique, real-world data — will keep pulling ahead. Not because their models are better. But because their data is.
English
