This paper examines the evolving structure and competition dynamics of the rapidly growing market for foundation models, with a focus on large language models (LLMs). We describe the technological characteristics that shape the AI industry and have given rise to fierce competition among the leading players. The paper analyzes the cost structure of foundation models, emphasizing the importance of key inputs such as computational resources, data, and talent, and identifies significant economies of scale and scope that may create a tendency towards greater market concentration in the future. We explore two concerns for competition, the risk of market tipping and the implications of vertical integration, and we evaluate policy remedies that aim to maintain a competitive landscape. Looking ahead to increasingly transformative AI systems, we discuss how market concentration could translate into unprecedented accumulation of power, highlighting the broader societal stakes of competition policy.
Authors
- Acknowledgements & Disclosure
- This is a revised version of a paper presented at the 79th Economic Policy panel meeting in Brussels in April 2024. We thank Susan Athey, Emma Bluemke, Claire Dennis, Avi Goldfarb, Aidan Kane, Pia Malaney, Sarah Myers West, Sanjay Patnaik, Nicholas Ritter, Max Schnidman, Eli Schrag, Rob Seamans and Joseph Stiglitz as well as our editors, Emilio Calvano and Giacomo Calzolari, our discussant Doh-Shin Jeon, and two anonymous referees for thoughtful comments and conversations. Any remaining errors are our own. Korinek gratefully acknowledges financial support from the Center for Innovation, Growth and Society of the Institute for New Economic Thinking (INET-CIGS, grant no INO21-00004). Jai Vipra gratefully acknowledges financial support as a Winter Fellow from the Centre for the Governance of AI. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research or of the other institutions that the authors are affiliated with. An earlier version of this paper was circulated under the title “Market Concentration Implications of Foundation Models: The Invisible Hand of ChatGPT.”
- DOI
- https://doi.org/10.3386/w33139
- Pages
- 31
- Published in
- United States of America
Table of Contents
- NBER WORKING PAPER SERIES 1
- CONCENTRATING INTELLIGENCE SCALING AND MARKET STRUCTURE IN ARTIFICIAL INTELLIGENCE 1
- Anton Korinek Jai Vipra 1
- Working Paper 33139 httpwww.nber.orgpapersw33139 1
- NATIONAL BUREAU OF ECONOMIC RESEARCH 1050 Massachusetts Avenue Cambridge MA 02138 November 2024 1
- Concentrating Intelligence Scaling and Market Structure in Artificial Intelligence Anton Korinek and Jai Vipra NBER Working Paper No. 33139 November 2024 JEL No. D43 K21 L4 L86 O33 2
- Anton Korinek Department of Economics University of Virginia Monroe Hall 246 248 McCormick Rd Charlottesville VA 22904 and NBER antonkorinek.com 2
- Jai Vipra Department of Science and Technology Studies Cornell University 415 Morrill Hall Ithaca NY 14853 jv474cornell.edu 2
- 1. Introduction 3
- 2. A snapshot of the market for generative AI 5
- OpenAI USA ChatGPT-4o-latest 2024-09-03 1340 6
- Google DeepMind USAUK Gemini-1.5-Pro-002 2024-09-24 1303 6
- Anthropic USA Claude 3.5 Sonnet New 2024-10-22 1286 6
- 3. Technological characteristics and market structure 12
- 4. Concentration Concerns 18
- 5. Conclusions 25
- Bibliography 27
- Economic Policy 27
- The Wall Street Journal 27
- Verfassungsblog 27
- Epoch 27
- Science 28
- The 28
- Verge 28
- The Yale Law Journal 29
- IMF Finance Development Magazine 29
- Wired 29
- Journal of Economic Literature 29
- IMF Finance Development 29
- Magazine 29
- Journal of Economic Literature 29
- Strategic 29
- Management Journal 29
- Google 30
- The GitHub Blog 30
- TechCrunch 30
- Federal Trade Commission 30
- The Washington Post 31
- The Information 31
- The Verge 31