Search

The DeepSeek Effect: Democratizing AI through Open-Source Ecosystems and Cost-Efficient Training

Author(s)

Hui Shi

Affiliation(s)

Guangzhou Huashang College, Guangzhou, Guangdong, China Graduate University of Mongolia, Ulaanbaatar,14200, Mongolia

Abstract

This study addresses key challenges in the democratization of artificial intelligence, including high computing power barriers, imbalanced industry penetration, and insufficient ecosystem sustainability. It proposes the theoretical framework of the "DeepSeek Effect," systematically explaining how open-source ecosystem collaboration and cost-efficient training paradigm innovation drive AI inclusivity. At the economic level, this study constructs a coordinated pricing model for the "algorithm-computing power-data" elements, validating the market reconstruction effects triggered by the near-zero marginal cost pricing strategy in an open-source ecosystem. At the practical level, based on localized case studies in China—such as the grassroots AI triage system in Jiangxi and the intelligent monitoring network for tea farmers in Yunnan—it reveals pathways for technology dissemination. For the first time, this study introduces a three-dimensional driving model of "technology cost reduction, ecosystem expansion, and scenario penetration," providing a theoretical basis for addressing the "Solow Paradox" in AI diffusion and offering quantitative references for policymaking on inclusive digital technology development.

Keywords

AI Popularization; DeepSeek Effect; Open Source Ecology; Developer Tools; Inclusive Applications

References

[1]Fedus, W., Zoph, B., & Shazeer, N. (2022). Switch Transformers: Scaling to Trillion Parameter Models. In Proceedings of the 39th International Conference on Machine Learning (pp. 234–267). PMLR. [2]Izacard, G. (2022). Atlas: Few-shot Learning with Retrieval-Augmented Generation. Advances in Neural Information Processing Systems (NeurIPS 2022). [3]Ma, L., Zhu, B. Q., & FYi, C. N. (2023). Network intelligence standards, open source and industry research. Telecommunications Science, 37(10), 12-21. [4]Zhang, Y., Wang, T., Yin, G., Yu, Y., & Huang, J. (2021). Big data in open-source ecosystems for intelligent software development. Big Data, 7(1), 94-106. [5]Zhang, X., Li, Y., & Wang, J. (2023). Research on AI intelligent computing infrastructure architecture and key technology analysis. Information and Communications Technology, 17(1), 56-63. [6]Li, H., Wang, S., & Chen, Y. (2023). China open source ecosystem map 2023: Artificial intelligence field. InfoQ Research Center. [7]Ma, L., Zhu, B. Q., & FYi, C. N. (2025). A study of the governance of knowledge sharing in open-source communities. European Journal of Innovation Management. [8]FYi, C. N., Ma, L., & Zhu, B. Q. (2025, March). A study of the governance of knowledge sharing in open-source communities. European Journal of Innovation Management. [9]Smith, J. (2022). AI in decision support systems: Enhancements in financial and agricultural monitoring. Journal of Artificial Intelligence Applications, 35(4), 210-225. [10]Chen, Y. L. (2024). Defining "Big Data Price Discrimination": Using Transparency of Differential Treatment as the Criterion. Economic Law Review, 44(2), 164-183.