In merger reviews, antitrust authorities often examine why a firm chooses to acquire a product, invest in another company, or form a strategic partnership rather than develop a product in-house, particularly when it appears to have the resources to do so.
The rapid growth of artificial intelligence (AI) makes the build-versus-buy question particularly important. AI markets are evolving quickly, and firms must decide whether to develop AI capabilities internally or access them through acquisitions and partnerships.
In the article, “Antitrust Implications of the Build vs. Buy Playbook in the AI Era,” CRA’s Yajing Jiang and Ishuwar Seetharam, together with co-author Arpita Patnaik, explore under which circumstances firms may find it most efficient or procompetitive to build their own AI foundation models versus acquire or partner with another company to access the models.
The authors focus on how rapid technological change and data’s role as an economic asset shape these decisions, with implications for competition among foundation models and their applications.
Key takeaways
- Building is not always better. Acquisitions and partnerships can help firms access AI capabilities, support innovation, and enter markets faster.
- AI presents distinct trade-offs. Fast-moving technology, computing costs, and the changing value of data can affect the advantages of developing models in-house.
- Competition safeguards matter. Antitrust reviews should weigh potential benefits against risks such as restricted market access and greater concentration, while considering safeguards such as non-exclusive partnerships and data interoperability.
Read the full article published in the ABA’s The Antitrust Source here.