Larry Ellison Says AI Race Will be Led by Those with Access to Private Enterprise Data

Oracle (NYSE: ORCL) co-founder and CTO Larry Ellison stated during the company's fiscal Q2 earnings call last month that the AI race is not about developing the most advanced model, but about what data those models are trained on.

Whether it's ChatGPT, Gemini, or Meta's Llama, most AI models learn from the same publicly accessible internet data, which Ellison warned is turning cutting-edge AI into a commodity, with little differentiation between offerings.

'All the large language models—OpenAI, Anthropic, Meta, Google, xAI—they're all trained on the same public data from the internet. So they're basically the same. That's why they're becoming commoditised so quickly,' Ellison explained, adding that this similarity across models will diminish the unique value propositions for users.

Private Data Is the Real Money

Ellison emphasised that securely accessing and working with proprietary enterprise data is the next big opportunity. He stated that future breakthroughs will come from leveraging private data.

The billionaire added that the second wave of AI will be more impactful than the current Generative AI boom, and Oracle believes it has an edge when it comes to accessing high-value corporate data.

How is Oracle Preparing to Capitalise on This New Opportunity?

Discussing Oracle's strategy to leverage private data, Ellison highlighted the company's AI Data Platform, which utilises methods such as Retrieval-Augmented Generation to enable real-time searches on proprietary data.

Oracle has also significantly increased its capital expenditure guidance to $50 billion (£36.5 billion) for 2026, up from the $35 billion (£25.6 billion) forecast just three months ago.

At Oracle AI World in October 2025, the company announced multiple infrastructure collaborations, including a 50,000-GPU supercluster powered by AMD's MI450 chips, scheduled for Q3 2026, and the OCI Zettascale10 supercomputer designed to connect hundreds of thousands of NVIDIA GPUs. By late 2025, Oracle's cloud backlog had surged, driven mainly by AI demand from enterprises.

Oracle Faces Intense Competition

While Ellison aims to position Oracle as a leader in the global AI landscape, he acknowledged increasing competition from major players such as Amazon Web Services, Microsoft Azure, and Google Cloud, all of which are aggressively enhancing their enterprise AI capabilities. Additionally, advances in synthetic data generation could reduce reliance on proprietary datasets, potentially reshaping the AI industry.

Overall, Ellison's insights highlight the pressing need for AI models to move beyond dependence on public data towards more personalised, secure, and enterprise-focused innovations. However, the key question remains whether Oracle's stronghold in enterprise databases will give it a significant advantage in this race or if the AI market will shift again before its infrastructure investments yield substantial returns.

Oracle's Stock Price Outlook

Recently, Oracle's stock increased by nearly 3% to $182.44 (£133.46) per share. Despite this uptick, the company has faced challenges over the past year, even amid the broader stock market rally driven by the AI boom. According to Wall Street analysts surveyed by TipRanks, the average 12-month target price for Oracle stock stands at $302.41 (£221.23), representing a potential 65% upside from its current level. The highest forecast among analysts predicts a target of $400 (£292.62) per share.

This recent positive trend in Oracle's stock price could also be linked to developments surrounding TikTok. Last week, TikTok secured its future in the US by finalising a deal to create a new entity, TikTok USDS Joint Venture. Oracle, Silver Lake, and Abu Dhabi-based MGX will each hold a 15% stake and serve as managing investors. TikTok announced that US user data and its algorithm will be hosted on Oracle's cloud infrastructure, with enhanced privacy and cybersecurity safeguards, helping to reassure regulators and users alike.