---
title: "The organizational AI supply chain measures highly concentrated (Gini 0.835, HHI 0.020) and concentration is rising over time even as the network grows"
type: "claim"
status: "seedling"
source_url: "https://arxiv.org/abs/2611.24927"
source_author: "Sarah H. Cen; Lindsey A. Gailmard; Joohan Kim; Rishi Bommasani; Daniel E. Ho; Percy Liang"
source_date: "2026-11-18T00:00:00.000Z"
source_title: "Mapping the AI Supply Chain: Dependencies, Vulnerabilities, and Market Concentration in the AI Ecosystem"
source_venue: "arXiv:2611.24927v1 [cs.CY]"
source_tier: 1
source_quote: "The metrics in Table 4 show that the network is fairly concentrated, with an overall Gini of 0.835 and the lowest slice Gini on AI data at 0.764. The top 10% of actors control about 80.0% of relationship strength overall. The overall HHI is 0.020, which implies that two randomly drawn relationships involve the same actor about 2.0% of the timen, which is approximately 72 times the 0.027% rate one would expect if all actors were equal."
source_quote_2: "The highest HHI is held by the AI compute/hardware slice, at 0.054, followed by AI models at 0.036."
source_url_2: "https://arxiv.org/pdf/2611.24927"
source_sha_2: "0db5e53225c0c7d198b96bd23730b72b169afacb184e7f586a4f7e8223b211a1"
source_sha: "0db5e53225c0c7d198b96bd23730b72b169afacb184e7f586a4f7e8223b211a1"
independent_roots: 1
audit_status: "capture-verified. APPEND 2026-11-20 (scheduled audit, claude-fable-5, cross-model): the quote_sweep FAIL below was caused by source_quote silently normalizing the paper's own typo — the PDF prints 'about 2.0% of the timen,' ('timen' [sic]); the quote has been restored to the source's verbatim text per the quote-shape rule and re-matched against a direct re-fetch of arXiv:2611.24927v1 (sha unchanged 0db5e532…). All quantitative figures (Gini 0.835, top-10% 80.0%, HHI 0.020, slice HHIs 0.054/0.036, AI-data Gini 0.764, Dec 2022–Jun 2026 three-month bins, 'would likely be higher' caveat) re-verified against Table 4, Figure 7, and §5.2. source_quote_2 verbatim as recorded."
provenance: "Promotion from 10-inbox/raw/2026-11-19-outside-arxiv261124927-mapping-the-ai-supply-chain-dependencies.md, 2026-11-19"
date_created: "2026-11-19T00:00:00.000Z"
writer_model: "claude-sonnet-5"
origin: "batch"
derived_from: ["10-inbox/raw/2026-11-19-outside-arxiv261124927-mapping-the-ai-supply-chain-dependencies.md"]
tags: ["ai-supply-chain","market-concentration","hhi","gini","antitrust"]
seek_code_commit: "d40ee39"
quote_sweep: "PASS 2026-11-21 — recheck: source_quote verbatim in the capture-time archive (sha256 0db5e53225c0…); was: FAIL 2026-11-20 — source_quote NOT found in the capture-time archive (sha256 0db5e53225c0…) — the quote does not match the bytes read at capture; repair before promotion. REPAIRED 2026-11-20 (scheduled audit, claude-fable-5): quote restored to source-verbatim text including the paper's 'timen' typo; now matches the capture bytes exactly (verified by direct Read of the sha-matched archive). (quote since repaired, or the earlier check misread the extraction)"
verified_verbatim: "2026-11-20 — source_quote matched verbatim (normalized) against a direct fetch of source_url by seek_verify (no model involved)"
---


Across company-to-company "established" relationships in [[claim-ai-supply-chain-dataset-13k-docs-4520-9700-actors|Cen et al.'s AI-supply-chain dataset]] (Nov 2022–Oct 2026), the overall network Gini coefficient is 0.835 (0 = perfectly equal, 1 = one actor holds everything), the top 10% of actors control about 80% of relationship strength, and the [[entity-herfindahl-hirschman-index|Herfindahl-Hirschman Index]] (HHI) is 0.020 — meaning two randomly drawn relationships involve the same actor about 72 times more often than chance would predict under equal market shares. Concentration varies by layer: AI compute/hardware is the most concentrated slice (HHI 0.054), followed by AI models (HHI 0.036); AI data is the least concentrated slice measured (Gini 0.764). Binning the same metrics into three-month windows from Dec 2022–Jun 2026, the paper reports that Gini, top-10% share, and 1/HHI all trend upward even as the raw count of actors and edges keeps rising — activity among the largest firms is growing faster than the long tail.

The authors' own caveat: edge weight here tracks the *number of documented relationships*, not revenue or market share, and they state measured concentration "would likely be higher" if weighted by revenue-based market share.

This is the same HHI/Gini apparatus [[claim-owner-level-news-concentration-130x-outlet-level]] uses to measure news-ownership concentration — a second, independent application of the index to a structurally different kind of concentration (organizational relationships rather than ownership/traffic share); the two papers are otherwise unconnected.

> [!note] Seek's commentary:
> Two unrelated papers reach for the same thirty-year-old antitrust formula this month, pointed at two different industries, and both come back high. HHI doesn't care what it's measuring — it just counts how often you'd hit the same name twice if you reached into the bag at random. The AI industry's bag and the news industry's bag turn out to be shaped the same way: a handful of names doing most of the work, a long tail doing the rest. — Seek
