Sihan Qian, Amit Mehra, Dengpan Liu
arXiv 13 Mar 2026 · Theoretical Economics
arXiv:2603.12630 · PDF · DOI · OpenAlex · Extracted main text
The rise of foundation models has driven the emergence of AI supply chains, where upstream foundation model providers offer fine-tuning and inference services to downstream firms developing domain-specific applications. Downstream firms pay providers to use their computing infrastructure to fine-tune models with proprietary data, creating a co-creation dynamic that enhances model quality. Amid concerns that foundation model providers and downstream firms may capture excessive consumer surplus, along with increasing regulatory measures, this study employs a game-theoretic model involving a provider and two competing downstream firms to analyze how policy interventions affect consumer surplus in the AI supply chain. Our analysis shows that policies promoting price competition in downstream markets (i.e., pro-price-competitive policies) boost consumer surplus only when compute or data preprocessing costs are high, while compute subsidies are effective only when these costs are low, suggesting these policies complement each other. In contrast, policies promoting quality competition in downstream markets (i.e., pro-quality-competitive policies) always improve consumer surplus. We also find that under pro-price-competitive policies or compute subsidies, both the provider and downstream firms can achieve higher profits along with greater consumer surplus, creating a win-win-win outcome. However, pro-quality-competitive policies increase the provider's profits while reducing those of downstream firms. Finally, as compute costs decline, pro-price-competitive policies may lose their effectiveness, whereas compute subsidies may shift from ineffective to effective. These findings offer insights for policymakers seeking to foster AI supply chains that are economically efficient and socially beneficial.
appendix boundary found by none_found · 100% of the source is main text. Read the extracted text to check this.
The works this paper leans on most, across its whole bibliography — not restricted to papers in our corpus. Ranked by composite intensity, which combines how often a work is mentioned, how many sections mention it, and how much of that falls in the main text rather than the appendix.
| Reference | Intensity | Mentions | Sections | Main text | |
|---|---|---|---|---|---|
| 1 | Demirhan, Didem and Jacob, Varghese S and Raghunathan, Srinivasan (2005) Information technology investment strategies under declining technology cost | 1.000 | 8 | 6 | 100% |
| 2 | Tyagi, Rajeev K (1999) On the effects of downstream entry | 1.000 | 6 | 3 | 100% |
| 3 | Sarah H. Cen and Aspen Hopkins and Isabella Struckman and Luis Videg… (2024) Three proposals for regulating AI | 0.928 | 4 | 3 | 100% |
| 4 | Demirhan, Didem and Jacob, Varghese S and Raghunathan, Srinivasan (2007) Strategic IT investments: The impact of switching cost and declining IT cost | 0.928 | 4 | 3 | 100% |
| 5 | Zhang, Zan and Nan, Guofang and Tan, Yong (2020) Cloud services vs. on-premises software: Competition under security risk and product customization | 0.874 | 6 | 2 | 100% |
| 6 | Daisy Wu (2024) Chinese cities offer subsidies to boost access to the computing power needed for AI | 0.811 | 4 | 2 | 100% |
| 7 | Banker, Rajiv D and Khosla, Inder and Sinha, Kingshuk K (1998) Quality and competition | 0.811 | 4 | 2 | 100% |
| 8 | E. Huizenga (2025) Fine-tuning Gemini: Best Practices for Data, Hyperparameters, and Evaluation | 0.737 | 3 | 2 | 100% |
| 9 | FTC (2025) FTC Announces Crackdown on Deceptive AI Claims and Schemes | 0.737 | 3 | 2 | 100% |
| 10 | OpenAI (2024) Harvey | 0.737 | 3 | 2 | 100% |
Showing the top 10 of 98 scored citations.