Dec.17_10:00_グローバル半導体エグゼクティブサミット

Next-generation AI and accelerated computing are revolutionizing how semiconductors are designed and built. In this keynote, NVIDIA General Manager (GM), Industrial and Computational Engineering, Tim Costa, will explore how AI supercomputing powers a new era of smart manufacturing—from AI driven chip design and physics-accurate simulation to autonomous robotics and photorealistic fab digital twins. We’ll show how Physical AI—intelligence embedded directly into hardware—combined with GPU accelerated workflows is compressing design cycles, boosting yields, and enabling fully optimized, adaptive fabs. Central to this transformation is the AI Factory—a modern datacenter architecture delivering the compute scale needed to unify semiconductor R&D, manufacturing, and datacenter innovation. These advances are reshaping the semiconductor value chain, driving unprecedented speed, precision, and competitiveness for the industry’s next decade.

The semiconductor industry is poised for rapid growth in the next 5 years, largely driven by increasing demand for AI.  As traditional scaling laws reach their physical and economic limits, the demand for compute systems that can scale efficiently across performance, power, and cost dimensions has never been greater. This talk explores emerging strategies to overcome these challenges, focusing on improved compute efficiency, power delivery, thermal management, and memory bandwidth.  Drawing on recent R&D insights and ecosystem collaborations, the session will highlight practical pathways to extend system-level scaling beyond transistor-level improvements.

Artificial intelligence is revolutionizing industries and reshaping daily life -- from autonomous vehicles and personalized healthcare to generative models and scientific breakthroughs. Memory is at the heart of this transformation – the critical catalysts enabling AI to access, store, and process vast datasets at unprecedented speed and scale. By driving sustainability and performance in memory technologies, the semiconductor industry must pave the way for AI to reach new heights while addressing technical innovation, global energy and environmental challenges. At Micron, we are optimizing chip and system architectures to deliver higher bandwidth, lower latency, and better power efficiency—enabling the next generation of AI applications. And to accelerate future AI progress, the next generation of advanced memory must continue to significantly optimize power efficiency, dramatically increase memory performance and embrace sustainable development and manufacturing approaches. This talk will explore how Micron meets these challenges through innovative memory technology and product design, intelligent R&D and smarter manufacturing.