Research · Living data · Updated 2026-07-17
Every public company mapped to the artificial-intelligence theme — from foundation-model builders to the picks-and-shovels supply chain.
"AI stocks" is the most-searched investing theme of the decade, and also the most abused — half the tickers marketed as AI plays barely touch the technology. NextTrack maps 530 public companies to the artificial-intelligence theme and its sub-themes, from the labs training foundation models to the chips, networking, power, and data-center supply chain that make them run. The count refreshes on every rebuild as new companies are classified, so this is a living map of the theme rather than a hand-picked top-ten.
| Metric | Live figure |
|---|---|
| NextTrack tracks | 530 AI-themed companies |
| Last updated | 2026-07-17 |
There is no single "AI sector" in the standard GICS taxonomy, which is exactly why a purpose-built theme map matters. NextTrack classifies a company into the AI theme when its business is materially exposed to artificial intelligence — not because a press release used the word once. The 530 companies here span several sub-themes: the model builders and AI-software applications; the semiconductor layer (accelerators, memory, networking silicon); the infrastructure layer (data centers, power, cooling, interconnect); and the enablers whose demand is being pulled forward by the AI build-out.
That layered view is the point. The theme is not one trade — it's a supply chain, and where you sit in it determines your risk.
The naive mistake is treating every name on an AI list as equivalent. They are not. A foundation-model company, a memory maker, and a utility selling power to a data center are all "AI-exposed," but they have wildly different margins, cyclicality, and valuations.
The useful lens is **purity versus size**. A small company whose entire revenue is AI is a high-beta, high-variance bet on the theme. A mega-cap where AI is a fast-growing slice of a diversified business is a lower-variance way to own the trend. NextTrack keeps both in the map and lets you sort by market cap and sub-theme so you can tell one from the other instead of lumping them together.
With 530 companies mapped, the AI theme in 2026 is defined less by the model labs — most of the marquee ones remain private — and more by the **public picks-and-shovels layer** that sells into them: accelerators and memory, the power and cooling required to run them, and the software companies embedding models into existing workflows. The full ranked list, sortable by market cap, sub-theme, and return, lives on the gated theme page. The biggest names by market cap anchor the list; the long tail is where the dispersion — and the risk — concentrates.
Theme membership is built from a curated taxonomy (`themes_taxonomy`) mapped to companies via a classification layer, then rolled up recursively so a parent theme like AI includes every company tagged to any of its sub-themes. Membership is reviewed as business mixes change, so a company that pivots into — or out of — AI is re-classified rather than frozen at its first tag. Explore the full interactive map on the [AI theme page](/themes/ai), or see the [complete theme directory](/themes).
Explore all market themes · Quantum computing stocks list · Nuclear energy stocks