Apply AI Agents to HBM and HBF Design and Verification to Boost Semiconductor R&D Efficiency
Build Research Environment Where One Engineer Directs Multiple AI Agents
Showcase Development Results on Nov. 11

Kim Jeongho, a KAIST professor known as the "father of high-bandwidth memory (HBM)," has launched research using artificial intelligence (AI) agents to strengthen South Korea's competitiveness in semiconductor technology. The aim is to deploy AI agents for complex semiconductor design, analysis and verification tasks to improve the efficiency and technical quality of next-generation memory development.

KAIST Professor Kim Jeongho

KAIST Professor Kim Jeongho

View original image

TERA Lab, led by Kim Jeongho, announced on Nov. 5 that it will hold the online "TERA AI Agent Day" on Nov. 11 to present AI agents developed by student researchers and examples of their application in research. TERA Lab is conducting research into applying AI agents to the design of HBM and high-bandwidth flash (HBF).


At the event on Nov. 11, around 15 student researchers will introduce AI agents they have applied to research tasks such as semiconductor design automation, paper writing and literature reviews. TERA Lab is also building a research environment in which each student develops and uses two or more AI agents.


The core of the research is a system in which one semiconductor engineer directs multiple specialized AI agents. AI agents specialized in different fields will divide complex tasks, ranging from circuit design and packaging to signal, power and thermal analysis. Researchers will verify the results and focus on key technical decisions.


TERA Lab researchers are applying AI agents to tasks including signal quality analysis and power-noise prediction for HBM and HBF, as well as optimizing the placement of through-silicon vias (TSVs), which vertically connect components inside semiconductors. They are also pursuing printed circuit board (PCB) design automation using a 3D electromagnetic field simulation tool from semiconductor design software company Synopsys.


The research is expected to reduce the time required for repetitive analysis and verification while improving design quality. In particular, it will help determine the potential of AI agents in developing next-generation semiconductors such as HBM and HBF, which require the integration of multiple technologies.




Professor Kim emphasized, "The trend of having AI agents take on and collaborate on these specialized tasks will bring fundamental changes to how semiconductor research and development is conducted." He added, "In the future, the competitiveness of engineers and developers will be determined not simply by their ability to use AI, but by their ability to design a range of specialized AI agents themselves and enable them to work together seamlessly."


This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing