Why Brain-Inspired Computing Could Redefine the Future of Energy-Efficient AI

Wilfred van der Wiel, coordinator of HYBRAIN, was recently interviewed for an article in which he shares his perspective on the current state of computing and the growing energy challenge driven by artificial intelligence.

In the interview, he explains why today’s computers, while extremely powerful, are fundamentally inefficient for AI-related tasks, and how brain-inspired computing offers a promising alternative by solving complex problems more efficiently at the hardware level. Drawing on HYBRAIN research, the interview highlights how neuromorphic approaches rethink the way computation, memory, and learning are combined, moving away from energy-intensive data shuttling toward systems that process information more like the human brain.

This contribution is highly relevant for HYBRAIN, as it clearly connects the project’s scientific goals with a pressing societal challenge: reducing the energy footprint of future AI technologies. Three key messages emerge from the interview:

• The biggest energy cost in modern computing comes from how data is moved and processed, not from the calculations themselves.
  
• The human brain remains the most efficient information processor we know, especially for tasks such as speech and pattern recognition.
  
• Brain-inspired hardware can enable more sustainable AI by performing these tasks directly and efficiently in the material, rather than relying on ever-larger data centres