Frontier stage
Keynote

Biochips for future AI computers

In person
  • Date
    10 July 2025
    Timeframe
    16:20 - 16:40 CEST
    Duration
    20 minutes
    Share this session

    AI algorithms were developed to emulate the human brain. These non- von Neumann type computational approaches are currently operated with silicon-based chips (GPUs, for example), but they can be very energy-intensive with significant implications for sustainability. In the future, we envision that AI computational devices could be made with living neural organoids, potentially making them more energy efficient. As part of a team effort at Johns Hopkins University to develop Organoid Intelligence (OI), we have designed, fabricated, and operated the first 3D integrated shell microelectrode neural electrode interfaces that mimic microscale-EEG caps. They biochips are scalable, mass-producible, biocompatible, and facilitate 3D spatiotemporal stimulation and recording with facile media change and long-term operation. We have achieved 3D spatiotemporal neuromodulation and created shells with microfluidic/electrical functionalities. We envision these biochips as the central building blocks of future energy-efficient AI computers and are developing these platforms to enable stimulus discrimination and reinforcement learning for real-world AI applications.

    Share this session
    • 17
      Days
      15
      Hours
      18
      Min
      05
      Sec