Archive · 10 Jan 2026
Larry Ellison Criticizes AI Models' Reliance on Public Data
Larry Ellison, the founder of Oracle, thinks that many AI models made by big companies like Google and OpenAI have a big problem.
They are all trained on the same public data from the internet, like websites and articles.
Because of this, these AI models give similar answers and don't offer much unique value.
Ellison believes that the future of AI should involve using private company data, like business records and customer information.
This way, AI can provide more useful and accurate results.
Oracle is working on AI tools for businesses that use their own private data to help with decision-making and operations.
Larry Ellison criticizes AI models for being trained on the same public data, making them similar.
He believes AI models need to use private company data for better insights and innovation.
Ellison made these comments during Oracle’s fiscal Q2, 2026 earnings call in December.
Oracle is focusing on building AI tools for enterprises that use private data.
The future of AI, according to Ellison, depends on who controls valuable data, not just who builds the biggest model.
- Who
- Larry Ellison, founder of Oracle
- What
- Criticism of AI models' reliance on public data
- When
- During Oracle’s fiscal Q2, 2026 earnings call in December
- Where
- Oracle’s earnings call
- Why
- To highlight the need for AI models to use private data for better insights
- Person
- Larry Ellison
- Company
- Oracle
- Issue
- AI models trained on same public data
- Solution
- Use private company data for AI training
- Impact
- Limited innovation and real value in AI
- Oracle's Focus
- AI tools for enterprises using private data