Scientists at Ohio State University have transformed dehydrated specimens of the edible mushroom Lentinula edodes into functioning memristors—electronic components that “remember” past current states—offering a paradigm-shifting approach to memory hardware that could usher in greener computing systems. The device achieved switching speeds of up to about 5,850 signals per second, operating with approximately 90 % accuracy under laboratory conditions.
The researchers grew the fungus, dehydrated the tissue to stabilise conductivity, wired the samples into bespoke circuits, and applied current across various frequencies and voltages. They found that distinct parts of the dried mycelium exhibited differing electrical characteristics, allowing the team to select optimal contact points for consistent performance. Senior scientist John LaRocco described their work as pushing memristive systems “to their limits.”
One of the motivations behind this research is the growing pressure in microelectronics to move beyond silicon and rare-earth-heavy materials, given resource constraints, rising fabrication costs and sustainability concerns. Fungi offer biodegradable, low-cost alternatives that reduce electronic waste and material extraction burdens. The mycelium-based memristors may be particularly suited for low-power, edge computing and specialised applications where ultra-high speed is less critical.
Despite the promising findings, significant obstacles remain before mushroom-based memory technology can be mainstreamed. Current performance still lags behind commercial semiconductor memory in terms of speed, density, and miniaturisation. The study acknowledges that entirely viable fungal memristors would require substantial shrinkage and integration into conventional electronics. Accuracy declines at higher frequencies, though the team found that adding multiple mushroom units in parallel can offset some performance loss.
The experimental devices achieved the switching frequency mark of ~5,850 Hz, a figure comparable to—but not yet exceeding—many conventional memristors. The fungal devices demonstrated robust memory behaviour, but reproducibility, long-term stability, environmental resilience and integration into chip-scale architectures remain open questions. The researchers note that larger mycelial systems could be used for aerospace, wearables or autonomous systems, but concede that these applications are preliminary.
Emerging trends in the field of bioelectronics suggest that alternative substrates—such as fungal mycelium, slime moulds and other organic systems—are attracting growing interest because of their adaptive, self-repairing and environmentally friendly properties. Mycelium networks already exhibit behaviour akin to neural signal transmission and have previously been shown to possess memristive behaviour, so the current study builds on earlier foundational research in unconventional computing.
Key players in this endeavour include LaRocco’s team at Ohio State and the supporting institutions such as the Honda Research Institute that provided backing for the work.
From a broader perspective, this research highlights the growing convergence of biology and electronics, where living or once-living systems are repurposed into functional hardware. The shift points to computing architectures that could be biodegradable, energy-efficient and customised for niche environments rather than aiming purely for raw speed. One of the co-authors emphasised that the resources needed to explore fungi-based computing could be as simple as “a compost heap and some homemade electronics” or as extensive as a “culturing factory with pre-made templates.”
Commercial adoption remains a long way off. The team acknowledges the need for advanced cultivation methods, reproducible manufacturing, shrinkage of device form-factor, and integration with existing silicon or hybrid platforms. Questions surrounding lifecycle, long-term durability, working in harsh environmental conditions and interoperability with standard CMOS systems all remain open. Nonetheless, this exploration opens a compelling pathway in the search for sustainable, brain-inspired computing platforms.