In 2022, we published a paper in Energy & Environmental Science showing that a small container of cyanobacteria, photosynthetic bacteria older than most things on this planet, could continuously power a commercial microprocessor using nothing but ambient light and water.

The idea sounds almost silly when you say it out loud. A chip, running computations, fed by the metabolic activity of tiny green organisms sitting on your windowsill. No batteries, no grid, no fossil fuels. Just sunlight, water, carbon dioxide, and a few billion years of evolutionary optimisation.

I remember how exciting it was when we first saw a stable current output that lasted days, then weeks, then months. There’s a peculiar feeling when an experiment works and you realise that the thing powering your measurement is alive, it’s growing, dividing, photosynthesising, doing all the things it has been doing since the Archean eon, except now it’s also keeping an ARM Cortex-M0+ running.

The device itself is humble. A small aluminium anode, a biofilm of Synechocystis, some wires. It sat on a windowsill in Cambridge. During the day, photosynthesis drove electrons out of water and through the circuit. At night the current continued, likely from respiratory breakdown of stored carbon. The thing just kept going.

Of course, the power output is tiny. We’re talking microwatts. You won’t charge your phone with this (yet…). But that’s not really the point. There are billions of devices in the Internet of Things that need only a tiny bit of energy, such as environmental sensors, weather stations, monitoring systems in remote locations where replacing batteries is impractical or impossible. For those, a self-sustaining, self-repairing biological power source starts to look less like science fiction and more like engineering.

The paper generated a lot of attention. The BBC covered it. Anton Petrov, a great science communicators on YouTube, made a video about it, which was surreal and wonderful to watch. The Register ran a piece with the preemptive clarification that no, it cannot play Doom. Which brings me to Reddit, where the most upvoted comment on every thread about the paper was, inevitably, some variation of “but can it run DOOM?” The internet is beautifully predictable. For the record: no, a cyanobacterium-powered ARM Cortex-M0+ cannot run DOOM. Not yet. Give evolution a few more billion years or give us millions in research funding and we can do that perhaps…

What I loved the most about this project was the collision of time scales. The ARM processor represents decades of semiconductor engineering and miniaturisation. The cyanobacteria represent billions of years of photosynthetic refinement. Bringing them together on a windowsill felt like a small, improbable meeting across deep time. If you’ve seen the Rick and Morty episode with the miniverse battery, where an entire civilisation exists solely to generate power for Rick’s car, you’ll appreciate the uncomfortable parallel. We built a tiny world of photosynthetic organisms, gave them light and water, and harvested their electrons to run our computations. The cyanobacteria, of course, have no idea. Or maybe they do, and somewhere inside a Synechocystis cell there’s an even smaller civilisation powering an even smaller microprocessor. It’s fractal slavery all the way down…

People were fascinated, and rightly so. Not because it solves the energy crisis (it doesn’t), but because it makes you reconsider what “technology” means. The most sophisticated solar energy converter on the planet is not a silicon panel, it’s a living cell. We just need to learn how to listen to it.

Read the paper: Powering a microprocessor by photosynthesis, Energy & Environmental Science (2022)