Speaker: Tobias Becker, NVIDIA
The Hardware of Intelligence: Application Specific Architectures for Next-Gen Agentic AI
Deploying agentic AI at scale requires more than just large models – it requires an architectural revolution. While AI agents promise breakthrough automation in coding, research, and many general office tasks, the computational demands of these agentic loops – defined by iterative reasoning and complex tool interaction – far exceed those of traditional Large Language Model (LLM) inference. Agentic systems therefore call for deep integration between model architecture and computer hardware. In this talk, we first dissect the core computational algorithms driving LLMs – the foundational technology behind AI agents. We show how heterogeneous, application-specific architectures are being engineered to maximise both performance and efficiency across the agentic stack. Finally, we offer a glimpse into the future where these agentic technologies can also revolutionise other domains, including industrial applications, robotics, and autonomous vehicles.
Speaker Bio
Tobias Becker is a Director in the NVIDIA LPU hardware team where he leads forward-looking explorations into new applications and accelerator architectures. He has a research and engineering background in AI, FPGAs, HPC and computer architecture. He received a PhD from Imperial College London and held previous positions at Groq, Maxeler, and Xilinx.