3 6 Ke·2025-11-25 05:59Core Insights - Carnegie Mellon University has unveiled a comprehensive dataset that outlines the flow of data, computing power, models, capital, and talent within the AI supply chain, highlighting the control points and potential bottlenecks in the industry [1][3][12] Group 1: AI Supply Chain Overview - The AI supply chain describes how AI models, data, and artifacts are produced upstream and adapted for use downstream [3][12] - The dataset was created by aggregating thousands of articles, press releases, and SEC filings, with ongoing weekly updates [3][6] - The emergence of the AI supply chain signifies a complex network of organizations involved in the development, deployment, and usage of AI systems [12][20] Group 2: Market Dynamics and Concentration - Market concentration in the upstream supply chain creates bottlenecks that could lead to cascading failures in products or services [5][9] - The dataset can reveal close financial relationships between AI organizations, including mutual investments and circular investments [6][9] - Understanding the dependencies within the supply chain is crucial for effective AI policy and governance, as it helps identify key inputs like computing power, data, and talent [9][22] Group 3: Industry Evolution and Maturity - Historically, AI and machine learning systems were developed internally by companies, but there has been a shift towards outsourcing certain processes, leading to the rise of specialized data service providers [18][20] - The specialization within the AI supply chain enhances efficiency and marks the industry's maturation, opening new avenues for innovation and competition [20][21] - A clear understanding of the AI supply chain is essential to discern who creates bottlenecks, who profits, and who bears the risks [21][22]