Forget AI how about a living computer? Could biocomputers help tackle IT’s huge emissions problem?
Writing in The Conversation, Lund University professor Heiner Linke argues that biological computers could dramatically cut energy use by working more slowly but in parallel—offering a radical shift from today’s power-hungry silicon processors.
Modern computing consumes vast amounts of energy, with data centres and devices accounting for 3% of global electricity demand.
As AI use grows, this figure is set to rise. But by rethinking how computers process information, scientists believe energy consumption could be slashed.
The idea is based on the Landauer limit, a principle established in 1961 that sets a theoretical minimum for the energy required to perform a computation.
Today’s processors, operating at billions of cycles per second, consume ten billion times this limit.
The faster a computer runs, the more energy it burns.
Biological systems offer a different approach. Instead of a single processor working at lightning speed, billions of smaller processors could perform tasks in parallel at a much slower pace—drastically reducing energy use while maintaining overall performance.
A 2023 study demonstrated that computers designed with massive parallel processing could function near the Landauer limit.
Nature already works this way: molecular machines inside cells perform highly efficient computations, consuming energy levels only slightly above this theoretical minimum.
The challenge is building reliable molecular networks that can process information collectively. While these biological systems won’t match the raw speed of silicon chips, they could provide a sustainable, ultra-low-energy alternative.
If computing is to become more energy efficient, slowing down might be the key. As Linke suggests, adopting nature’s approach could help curb the environmental impact of our digital world.