By Hung Ou Yang
With the continuous advancement of artificial intelligence ("AI"), the legal fields it affects have become increasingly diverse. AI will inevitably involve competition law issues related to both restraints on competition and unfair competition. In July 2025, the Taiwanese Fair Trade Commission ("FTC") released the "Explanatory Materials on Soliciting Public Opinions Regarding Competition Law Issues Concerning Generative Artificial Intelligence," which provides a preliminary overview of potential competition issues arising in the industry. The FTC further outlines the market structure and characteristics of generative AI, as well as the types of conduct subject to competition law regulation, demonstrating the high level of attention the competent competition authority in Taiwan has devoted to AI. According to the FTC's press release, the explanatory materials are intended not only to solicit public comments, but also to serve as an important foundation for future enforcement. Its significance is therefore self-evident.
Unilateral abuses of market power have drawn significant attention from the competent authority
According to the explanatory materials released by the FTC, in the generative AI industry, unilateral abuses of market power may be reflected in the following four major aspects:
(1) Access to computing resources
The term "computing resources" here mainly refers to AI semiconductor chips, with GPUs being the most critical. The primary market player is NVIDIA. However, since the construction of GPUs, servers, and other hardware involves significant costs, developers may not be able to afford such high expenses on their own. As a result, renting computing power through cloud services has become an alternative option. Therefore, providers of cloud services are also considered players in the industry who possess computing resources.
The FTC is concerned that, if entities controlling computing resources impose restrictions on new entrants’ access to such resources (for instance, by refusing to sell GPUs separately and requiring simultaneous purchases of other hardware), or if they give preferential access to their own partnered model developers, such practices could create entry barriers for other model developers.
(2) Restrictions on Access to Data
The FTC points out that large technology firms, while providing digital services, may also simultaneously engage in data collection. Since data is an essential input for AI model development, large digital platforms are naturally positioned to gain a competitive advantage in this regard. New AI model developers, lacking access to data of comparable scale and quality, may find it difficult to compete. In addition, issues such as whether consumers are allowed to transfer their data across products or services, or whether restrictions on competitors' access to specific datasets could affect the provision of AI services, are also of concern.
(3) Cloud Service Platforms
As major cloud service providers strive to establish their own service ecosystems, competition law focuses on whether difficulties in switching between ecosystems may create concerns. In addition, the FTC highlights the issue of tying arrangements: if cloud service providers condition access to cloud services on the use of their AI models, or bundle such services with other existing digital offerings, this may adversely affect competition in the relevant markets.
(4) Foundation Models
The FTC identifies three potential situations in which firms with significant market power in foundation models may abuse that power: (a) Requiring that end products (such as smartphones or automobiles) be equipped exclusively with their own foundation models; (b) Prioritizing the display of their own products or services in the outputs generated by the foundation model; and (c) Through the terms of use of the foundation model, requiring users to grant access to internal data generated from their use of AI services, thereby strengthening their ability to acquire data.
Mergers, Collaborative Conduct, and Misleading Advertising Are Also of Concern
Regarding mergers, the past practices traditionally would focus primarily on horizontal mergers between competitors, as the impact of increased market share is more immediately observable. However, in the generative artificial intelligence industry, the FTC notes that, in addition to horizontal mergers requiring careful scrutiny, vertical mergers can enhance market power in upstream or downstream markets. Diversified mergers may also create the ability or incentive to engage in tying and bundling strategies. Therefore, even merger types that do not increase market share remain within the scope of regulatory concern.
In sum, although the materials released by the FTC provide only a preliminary overview of competition law issues related to generative AI, it is evident that the content touches on nearly all major provisions of the Taiwanese Fair Trade Act. Therefore, businesses planning to implement AI applications should incorporate professional assessments from a competition law perspective. In addition, companies encountering competition law issues during AI implementation may also take advantage of the Taiwanese authorities’ public consultation process to appropriately express their views and contribute to the formation of future policies.
AUTHOR: Hung Ou Yang
Managing Partner
Taipei
+886-2-2707-9976
[email protected]
Copyright: Brain Trust International Law Firm
Disclaimer: While every effort has been made to ensure the accuracy of this publication, it is not intended to provide legal advice as individual situations will differ and should be discussed with an expert and/or lawyer. For specific technical or legal advice on the information provided and related topics, please contact the author.
撥打電話