At a Leuven briefing, imec's leadership warned that AI's hunger for compute is on a collision course with planetary limits. Their answer: squeeze far more performance per watt by pushing Moore's Law, going 3D, and lighting up chips with photonics.

At the launch of TU/e's Casimir Institute, speakers warned of AI's rising electricity footprint. From chip designers to policymakers, all agreed: AI's growth curve will collide with the planet's energy limits unless solutions are found.

Eindhoven-based Euclyd inference architecture promises the lowest power and cost per token for the next wave of agentic AI. The CRAFTWERK system delivers 1.024 exaFLOPS with a projected 100× improvement in power efficiency.

Mathieu Baltussen, a PhD candidate at Radboud University, has developed a prototype chemical computer using the Formose reaction. This innovative approach leverages complex chemical reaction networks for computing tasks with significantly less energy than traditional digital computers.

Professor Martijn Heck from TU Eindhoven explains how photonics can revolutionize AI energy efficiency by transferring data using light instead of electricity, enabling faster processing with much lower energy consumption.

Professor Christian Mayr from Dresden University explains how AI can draw inspiration from the human brain's remarkable 20-watt energy efficiency to revolutionize artificial intelligence systems.

The growing energy consumption of data centers makes it essential to manage the grid intelligently and make data centers more sustainable. Dutch Data Center Association explores connecting to high-voltage grid.

New study reveals a Gemini AI query uses less energy than nine seconds of TV - but scale still matters. Google offers the most detailed picture to date of AI's environmental impact.

Lumai's 3D optical computing technology promises to slash AI chip energy consumption by 90% while boosting performance by 50x, using light to process information with minimal energy.

CEO Dr. Hamed Sadeghian explains how Nearfield Instruments aims to help solve AI's growing energy problem through advanced semiconductor metrology.

Even if you want to avoid using AI, you probably can't. The rising energy costs are worrying, an MIT Technology Review roundtable showed.

Neuromorphic computing is maturing into a viable tool for energy-efficient, real-time processing, opening up opportunities for businesses.