Investors point to Eli Lilly’s decade-spanning metabolic data and AI build-out as a rival to today’s AI chip leaders

For most investors, the race to become the world’s most valuable company is often framed as a showdown between AI chip leaders and a handful of other “AI-first” businesses. But a growing camp of market observers is making a different bet, according to a recent discussion highlighted by 24/7 Wall St. The claim is that the biggest AI winner in the next few years may not be a company selling chips, but a pharmaceutical manufacturer sitting on decades of disease-related information that is harder to replicate than model weights alone.

The argument centers on Eli Lilly (NYSE: LLY), an Indiana-based company founded in the 19th century. The thesis, attributed in the post to investor Jordi Visser on The Pomp Podcast, is not that Lilly will outcompete NVIDIA (NASDAQ: NVDA) by building the same kind of compute platform. Instead, Visser’s view is that Lilly has a credible path to becoming the world’s largest company within about five years because it combines a growing AI capability with proprietary healthcare data and the commercial momentum of its obesity and diabetes treatments.

The post places Lilly’s current financial strength at the center of the case, noting that the company’s market capitalization has reached about $1.08 trillion after the success of GLP-1 drugs. GLP-1, or glucagon-like peptide-1, is a class of hormone-based therapies that help regulate appetite and blood sugar, and Lilly’s branded versions have helped shift the company’s earnings profile and investor narrative.

From there, the post pivots to AI infrastructure and data. It says Lilly is building a private AI infrastructure reportedly built around roughly 1,000 NVIDIA Blackwell GPUs. Blackwell refers to NVIDIA’s latest generation of AI accelerator chips, designed to train and run large-scale machine learning models more efficiently. In the post’s framing, assembling that kind of internal compute capacity is meant to support AI use cases across drug development and other operations.

Visser’s broader point, as summarized in the article, is that the hardest-to-copy asset may be Lilly’s 150 years of proprietary metabolic disease data, including information tied to diabetes, obesity, and related metabolic conditions. The post suggests that this dataset, accumulated over decades, could provide a foundation for more effective AI applications than what’s possible with generic data alone. That is, the differentiator is not merely access to GPUs, but the ability to apply AI to domain-specific information accumulated through routine clinical care and product history.

The post also treats Lilly’s AI spending as potentially more than isolated experiments. It argues that, taken together, Lilly’s AI initiatives make the company look less like a purely traditional drugmaker and more like an organization investing in large-scale AI applications. Even so, the article does not provide detailed documentation of the infrastructure, the specific use cases, or the performance outcomes that investors might look for in filings, earnings materials, or technical disclosures.

The discussion lands in a familiar debate about where value is created in the AI economy. Semiconductor companies have benefited from demand for training and inference hardware, while software and data platforms often capture the next layer of economics if they can translate compute into durable product advantage. Lilly, in this telling, is trying to link both layers by using AI internally while leveraging a steady pipeline of metabolic disease demand.

What is not clear from the post is the extent of Lilly’s AI roadmap beyond the reported GPU scale, including whether those systems are focused on specific drug targets, operational automation, or broader model development. The article also does not cite formal statements confirming the GPU count, the architecture, or any quantified results from the AI build-out. Investors and analysts will likely want to see more transparent updates in company presentations, regulatory filings, or earnings call discussion to validate how AI is being translated into measurable outcomes.