← News

AI securitySources: arXiv · 19 Nov 2026

Researchers release large-scale dataset mapping the AI supply chain

An arXiv paper introduces a pipeline that catalogs how models, data, chips, infrastructure, capital and talent flow between organizations, with the dataset publicly available.

In plain words

A research team built a large dataset describing which organizations contribute to building and deploying AI, and how they depend on each other. Think of it as a map of the AI industry's supply lines, showing who provides models, data, chips, money and people to whom. The dataset is free to download and explore online.

What was announced

The authors argue that while the layers of the AI stack are relatively well understood, the relationships between actors and components across those layers remain sparsely documented. Their pipeline produces a multilayered dataset of actors and relationships, annotated with dates, descriptions, types and tags, cataloging how models, data, chips, infrastructure, capital and talent move between organizations. They say this enables examination of industry relationships and of concentration in AI development.

Illustrative analyses in the paper include identifying the most central actors, measuring network concentration, and finding communities and chokepoints. The authors also flag motivations such as concerns about cascading failures and fragmented accountability. They caution that the dataset, while fairly comprehensive, is necessarily incomplete: public documents omit private deals, and coverage depends on how the search was designed. The data is hosted on Hugging Face, with an interactive explorer at aisupplychains.org.

What it does not show

  • The evidence is the paper's abstract only; the underlying methodology, data quality and findings could not be independently verified from this excerpt.
  • The authors themselves state the dataset is incomplete because public documents omit private deals and coverage is shaped by search design.

For security and IT teams

Access
The dataset is publicly available on Hugging Face and can be explored interactively at aisupplychains.org.

Sources

Source-support review completed 20 Nov 2026. This story was drafted with AI assistance and checked against retained source passages. Automated checking is not independent verification.

Coverage: arXiv: abstract.

Updates and corrections

No corrections so far.

First published 20 Nov 2026

New to the subject? AI from the Ground Up starts from the beginning.