McKinsey – Generative AI: the next S-curve for the semiconductor industry

The semiconductor industry is experiencing a surge in demand for computational power due to the rise of generative AI applications. Leaders are investing in expanding data centers and semiconductor fabs to meet the evolving needs of the market driven by gen AI. McKinsey has developed scenarios for the B2B and B2C markets, projecting a significant increase in wafer demand. The article discusses the estimated demand for high-performance components and corresponding fabs needed to supply them. There are two main types of gen AI applications: B2C and B2B, with varying compute demands for training and inference phases. To help semiconductor leaders navigate these demands, McKinsey outlines six use case archetypes for B2B compute demand and their value creation. B2C applications are expected to account for the majority of gen AI compute demand, with B2B use cases making up the remainder. By understanding the cost to serve and value creation metrics for each use case archetype, decision-makers can make informed choices in adopting gen AI technologies.