All LLMs Are Trained on the Same Data’: Oracle Chief Flags AI’s Biggest Challenge
During Oracle's fiscal Q2 2026 earnings call in December, he warned that the shared foundation was rapidly turning cutting-edge AI into a commodity product with razor-thin differentiation.
Topics
News
- Meta Signals Shift Toward Paid Features With Subscription Tests Across Its Apps
- All LLMs Are Trained on the Same Data’: Oracle Chief Flags AI’s Biggest Challenge
- Concern Grows as ChatGPT References Musk’s Grokipedia in Factual Answers
- TII and Qualcomm Bet on Edge AI to Move Autonomous Systems Out of the Lab
- Musk Predicts AI Will Outpace Human Intelligence by 2026
- UAE Companies Double Down on AI, but Gaps Persist: Report
[Image source: Chetan Jha/MITSMR Middle East]
As the race for artificial intelligence dominance continues, Oracle co-founder and CTO Larry Ellison has highlighted data as the next big challenge IT players will need to overcome.
Every major AI model—from ChatGPT to Grok—is being trained essentially on the same public data. During Oracle’s fiscal Q2 2026 earnings call in December, he warned that the shared foundation was rapidly turning cutting-edge AI into a commodity product with razor-thin differentiation.
”All the large language models—OpenAI, Anthropic, Meta, Google, xAI—they’re all trained on the same data. It’s all public data from the internet,” Ellison said. “So they’re all basically the same. And that’s why they’re becoming commoditized so quickly.”
The Next Bet?
Ellison further hints at what can be the next big thing for the IT players. He notes that the next gold rush won’t be building better foundational models but about data partnerships where AI will work with proprietary enterprise data while keeping it secure.
“The future lies in leveraging private enterprise data,” Ellison highlighted, suggesting this could potentially be more valuable than the current boom of GPUs, data centers, and public-model infrastructure.
Oracle has taken steps towards this prediction. The company now projects roughly $50 billion in capital expenditures for the full fiscal year—up from $35 billion estimated just three months earlier. Its recent announcements include AMD MI450 chips powered by a 50,000-GPU supercluster and the OCI Zettascale10 supercomputer, which links hundreds of thousands of Nvidia GPUs.
It’s a natural progression since most high-value private data already lives in Oracle databases. For instance, Oracle’s AI Data Platform, using techniques such as retrieval-augmented generation, enables models to query private information in real time without compromising security.
Ellison’s thesis still faces strong competition from Amazon Web Services, Microsoft Azure, and Google Cloud, which are racing to build similar enterprise AI capabilities. Additionally, synthetic data generation could also reduce reliance on exclusive proprietary datasets.
The global AI market will soar from $189 billion in 2023 to $4.8 trillion by 2033 – a 25-fold increase, predicts a new UN Trade and Development report.
