Mapping the AI Supply Chain: Dependencies, Vulnerabilities, and Market Concentration in the AI Ecosystem
Organizations: Carnegie Mellon University · Stanford University
Abstract
AI systems are built across many organizations that contribute models, data, compute, capital, talent, and other assets. While the layers of the AI stack are relatively well understood, the relationships between actors and components across layers remains sparsely documented. In this work, we focus on the organizations that make up the AI supply chain: the complex network of actors that contribute to the development and deployment of AI. We motivate the examination of the organizational AI supply chain---highlighting concerns about concentration, power, chokepoints, cascading failures, fragmented accountability, and risk---then describe our pipeline for mapping the AI supply chain. Our pipeline yields our main contribution: a multilayered, large-scale dataset of actors and relationships in the AI ecosystem, annotated with dates, descriptions, types, tags, and more. This work takes a step toward systematically mapping and measuring the AI supply chain and its properties, cataloging how models, data, chips, infrastructure, capital, and even talent flow between organizations. This data enable critical examinations of industry relationships and yield insights into the political economy of AI development and concentration. We provide several illustrative analyses, including studying the most central actors, analyzing network concentration, and identifying communities and chokepoints. While our dataset is fairly comprehensive, it is also necessarily incomplete: public documents omit private deals, and coverage is shaped by search design. Together, the approach, data, and findings aim to serve as a resource for studying the AI ecosystem. The dataset is publicly available at https://huggingface.co/datasets/sarahcen/ai-supply-chain and can be explored interactively at https://aisupplychains.org/.
Figures & tables
| Field | Description | Possible values |
|---|---|---|
| link | Source document URL | text (URL) |
| article_title | Source page title | text |
| document_date | Document/article date | date or missing |
| actor_1_name , actor_2_name | Actor names | text |
| actor_1_type , actor_2_type | Actor types | 11 labels. Table 16 contains the full set of labels, including as examples COMPANY , GOVERNMENT_BODY , and ACADEMIC_INSTITUTION . |
| relationship_description | One-sentence description | text |
| Slice | Num Nodes | Gini | Top 10% | HHI | |
|---|---|---|---|---|---|
| Overall | 3,677 | 0.835 | 0.800 | 0.020 | 51.0 |
| Investment | 2,020 | 0.780 | 0.728 | 0.012 | 83.1 |
| Provision | 1,705 | 0.831 | 0.792 | 0.025 | 39.7 |
| AI compute / hardware | 740 | 0.881 | 0.855 | 0.054 | 18.4 |
| AI data | 559 | 0.764 | 0.711 | 0.026 | 38.2 |
| AI models | 1,204 | 0.873 | 0.846 | 0.036 | 27.6 |
| Pair | Months | Investment | Provision | Partnership |
|---|---|---|---|---|
| Microsoft-OpenAI | 48 | 48 | 39 | 47 |
| Anthropic-Google | 44 | 40 | 34 | 24 |
| Meta-NVIDIA | 34 | 0 | 25 | 0 |
| X-xAI | 33 | 0 | 30 | 0 |
| Amazon-NVIDIA | 32 | 0 | 22 | 0 |
| China-United States | 30 | 0 | 0 | 0 |
Appendix figures & tables22 assets
Supplementary material from the paper’s appendix.
Appendix
| Rank | Domain | Count | Share (%) |
|---|---|---|---|
| 1 | finance.yahoo.com | 5,950 | 19.2 |
| 2 | techcrunch.com | 3,453 | 11.2 |
| 3 | linkedin.com | 2,861 | 9.3 |
| 4 | venturebeat.com | 1,233 | 4.0 |
| 5 | google.com | 593 | 1.9 |
| 6 | ca.finance.yahoo.com | 534 | 1.7 |
| Field | Description | Possible values |
|---|---|---|
| id | Primary key | integer |
| title | Result title from SerpAPI | text |
| link | Result URL | text (URL) |
| snippet | Result snippet | text |
| source | Publisher / source label | text |
| date | Parsed article date | datetime, or missing |
| Field | Description | Possible values |
|---|---|---|
| id | Primary key | integer |
| dedupe_key | Unique key of stripped URL | SHA-256 hex |
| link | Page URL | text (URL) |
| date | Article date string from SerpAPI | text, or missing |
| snippet | Search snippet carried to the scrape row | text |
| source | Publisher / source label | text |
| Field | Description | Possible values |
|---|---|---|
| id | Primary key | integer |
| dedupe_key | Key to gsearch_link_scrape.dedupe_key | SHA-256 hex |
| link | Source document URL | text (URL) |
| article_title | Title copied from scrape row | text |
| document_date | Parsed article date | date, or missing |
| actor_1_name , actor_2_name | Actor names | text |
| Quantity | Count |
| Total rows in gsearch_relationship_results | 314,460 |
| After AI + classification + date filter | 200,916 |
| Kept (after noise filter) | 100,079 |
| Discarded (total) | 214,381 |
| Undated dropped | 33,914 |
| Before start dropped | 7,435 |
| Actor grouping | Unique actors | Undirected pairs |
|---|---|---|
| Raw (pre-grouping strings) | 15,406 | 29,053 |
| Fuzzy actor grouping | 12,907 | 25,617 |
| Coarse actor grouping (pre-frequency filter) | 9,699 | 20,228 |
| Graph | Actor grouping | Edge norm. | Time grain | Nodes | Edges |
|---|---|---|---|---|---|
| Fuzzy | Raw | — | 5,602 | 89,297 | |
| Fuzzy | Indicator | Month | 5,602 | 39,565 | |
| Fuzzy | Indicator | Week | 5,602 | 43,767 | |
| Coarse | Raw | — | 4,520 | 92,197 | |
| Coarse | Indicator | Month | 4,520 | 37,472 | |
| Coarse | Indicator | Week | 4,520 | 42,449 |
| Rank | Actor | Total degree |
|---|---|---|
| 1 | 1,535 | |
| 2 | NVIDIA | 1,249 |
| 3 | Microsoft | 1,139 |
| 4 | OpenAI | 1,138 |
| 5 | Anthropic | 864 |
| 6 | Meta | 600 |
| AI category | Nodes | Top 5 (degree) |
|---|---|---|
| AI_PARTNERSHIP | 2,337 | Google (953), NVIDIA (812), Microsoft (700), OpenAI (579), Anthropic (433) |
| AI_FUNDING | 2,249 | NVIDIA (238), Google (197), OpenAI (164), Anthropic (155), Microsoft (135) |
| AI_SOFTWARE | 966 | NVIDIA (202), Google (126), Microsoft (106), OpenAI (95), IBM (76) |
| AI_INFRASTRUCTURE | 991 | NVIDIA (450), Google (218), Microsoft (193), Amazon (138), Meta (104) |
| AI_MODELS | 1,467 | Google (486), Anthropic (483), OpenAI (462), NVIDIA (264), Meta (237) |
| AI_APPLICATIONS | 1,585 | Google (434), OpenAI (311), Microsoft (304), Anthropic (198), NVIDIA (130) |
| Field | Description | Possible values |
|---|---|---|
| link | Source document URL | Text (URL) |
| article_title | Title associated with the source page | Text |
| document_date | Parsed document/article date, when available | Date, or missing |
| actor_1_name | Name of actor 1 | Text |
| actor_1_type | Entity type of actor 1 | COMPANY , NONPROFIT_OR_CIVIL_SOCIETY , ACADEMIC_INSTITUTION , COALITION_OR_CONSORTIUM , PERSON , GOVERNMENT_BODY , GEOPOLITICAL_ENTITY_OR_COUNTRY , INTERGOVERNMENTAL_OR_SUPRANATIONAL , OTHER_ACTOR , NOT_AN_ACTOR , UNKNOWN_OR_UNCLEAR |
| actor_2_name | Name of actor 2 | Text |