Beyond the Monolithic Chip: Scaling the Next Generation of AI Compute
Frost & Sullivan Podcasts
As AI and high-performance computing push the limits of conventional processor scaling, semiconductor architecture is becoming increasingly modular.
In this episode of Frost & Sullivan’s Growth Podcast, we examine how chiplets, high-bandwidth memory, Universal Chiplet Interconnect Express, and advanced 2.5D and 3D packaging are reshaping the next generation of AI and HPC systems.
We explore why semiconductor performance is increasingly moving to the system level, how HBM3E and HBM4 are addressing AI’s memory requirements, and why UCIe could accelerate interoperable, multi-vendor chiplet ecosystems. We also examine the growing strategic importance of hybrid bonding, heterogeneous integration, co-packaged optics, and optical chiplets as compute architectures continue to scale.
👉 Why are AI and HPC accelerating the transition from monolithic processors toward chiplet-based architectures?
👉 How will UCIe, HBM4, and advanced packaging influence the next generation of AI compute?
👉 Where are the strongest growth opportunities emerging across the semiconductor and advanced packaging ecosystem?
Designed for semiconductor companies, foundries, integrated device manufacturers, fabless companies, OSAT providers, EDA companies, hyperscalers, technology investors, and innovation leaders, this episode examines the architectural and ecosystem shifts shaping the future of AI and high-performance computing.
Explore related research:
https://www.frost.com | https://store.frost.com
Connect with our team:
hello@frost.com
Loading