Tech • AI • Robotics • Game
BIOCOMPUTING: These fungi can control robots… and this is only the beginning!
7/10TL;DR
Biocomputing is emerging as a low-energy alternative to conventional AI hardware, using living cells and organisms such as human neurons and fungi to process information, sense environments and control machines.
KEY POINTS
Energy pressure is driving interest in living computers
Modern AI systems require vast amounts of electricity, pushing major technology companies to secure dedicated power supplies, including nuclear output comparable to the consumption of 800,000 U.S. homes. That has intensified interest in computing systems that can perform useful tasks with radically lower energy demand. The human brain is a powerful benchmark, operating on roughly 20 watts while handling perception, adaptation and decision-making that remain difficult for many machine systems.
Wetware connects biological signals to machines
Wetware refers to systems that combine living organisms with electronics so biological electrical activity can be read, interpreted and turned into actions. The core requirement is that the organism produce stable, ordered signals rather than random noise. In practice, that can turn a biological response into a computing input, a sensor reading or a command for a machine.
Biological sensing is already in use
Some water networks already rely on living organisms as environmental sentinels. In Warsaw, freshwater mussels are used to detect contaminants, pesticides and heavy metals before water enters the municipal supply, while St. Petersburg has monitored water conditions through changes in crayfish heart activity. In both cases, stress responses are translated into electrical signals that can trigger protective actions such as closing valves.
Human neurons have reached commercial experimentation
The field has moved beyond theory into early commercial services described as wetware-as-a-service. These platforms offer access to lab-grown neural tissue derived from differentiated stem cells, allowing researchers to test computation and machine training on living neurons rather than Nvidia-style GPUs. The approach avoids harvesting tissue from human bodies, but it raises ethical concerns and practical hurdles around the care and stability of fragile neural cultures.
Cortical Labs has showcased neuron-based game control
One of the best-known demonstrations came from Cortical Labs in Australia, where a neuron-based system was used to play Doom. The setup relied on living neurons in a dish linked to a computer interface, illustrating how biological cells can be looped into a digital environment and learn to produce usable responses. The experiment became a vivid example of how wetware can blur the line between computation and living tissue.
Fungi may be better suited than neurons for many tasks
Researchers are increasingly exploring fungi because they are more robust than human neurons and still generate organized electrical impulses. Andrew Adamatzky of the Unconventional Computing Laboratory in Bristol has published extensively on fungal sensing and computing, including work arguing that fungal electrical patterns are species-specific, repeatable and information-rich. That makes fungi attractive as biological components for durable, low-maintenance computing systems.
Fungal robots and memory devices are advancing
In 2024, research published in Science Robotics described robots controlled by fungi, including a wheeled machine and another with five mechanical arms, both driven by Pleurotus eryngii. In 2025, another team reported a MemResistor made with shiitake mycelium, a component intended to combine traits of RAM and SSD into fast, non-volatile memory. Such devices could support more energy-efficient hardware while reducing dependence on rare earth materials and opening niche uses in harsh environments, including space, because mycelium can resist radiation.
The biological scale of fungi expands the vision
A fungal network is not limited to a small lab sample. The mycelium of Armillaria ostoyae in Oregon’s Malheur National Forest spans nearly 10 km², making it the largest known living organism on Earth, with estimates of 7,500 to 35,000 tons and an age of 2,400 to 8,650 years. That scale has fueled speculation that fungal systems could one day act as dense, distributed sensing networks capable of collecting and storing environmental information across vast three-dimensional spaces.
CONCLUSION
Biocomputing remains experimental, but it is moving from curiosity toward practical research in sensing, robotics and memory. The central question is no longer whether living systems can compute, but how society will use that capability as demand grows for cleaner and more efficient technology.
Ask a question