Published: 03 April 2024
Summary
GenAI’s future performance, cost and energy efficiency will be defined by semiconductor innovations and transforming supply chains. Product leaders reliant on chips for AI-related products and services must aggressively pursue supplier strategies to meet their product and business plans.
Included in Full Research
Overview
Implications
Because off-the-shelf chips will not meet future generative AI (GenAI) compute requirements, more large companies offering AI-related products and services will work directly with foundries, bypassing traditional fabless companies.
As monolithic system-on-chip (SoC) gets more expensive and less energy-efficient, semiconductor chiplets will drive future GenAI products’ innovation, implying a new supply chain, new design architectures and different testing methodologies for AI product leaders to address.
Traditional design for manufacturing approaches will not meet the power-performance metrics required for GenAI processing, which requires innovative approaches across the semiconductor value chain, starting with design and electronic design automation (EDA).
Conventional materials used
Clients can log in to view the entire
document.