This AI-Driven Bio-Processor Runs on Brain Cells!

Forget silicon chips—the future of AI might just involve a dash of dopamine and a sprinkle of human neurons. Swiss tech firm FinalSpark is breaking all the rules with their latest processor, which uses living components—specifically, mini brain-like organoids built from human brain cells—to deliver mind-blowing energy efficiency.

The system connects four organoids per unit, each with a cluster of human neurons linked to eight electrodes. These electrodes provide the necessary electrical stimulation to get the neurons firing, while a bit of dopamine helps keep the party going. The neurons communicate through what can only be described as a “bio-language” made up of chemical signals. It’s still unclear exactly how this will allow users to perform computational tasks, but that’s part of the experimental charm.

For now, FinalSpark is working with university researchers to observe and fine-tune the bio-computers. One of the big unknowns is what kind of tasks these living processors might excel at, and how they’ll stack up against our trusty silicon-based machines. The company is already brainstorming how to combine these bio-processors with AI systems, which are notorious energy hogs, to make things a little more eco-friendly.

The current organoids contain about 10,000 neurons and measure just half a millimeter across. Created from human-induced pluripotent stem cells (iPSCs) derived from skin tissue, these neuron clusters are kept alive at a toasty 37°C, shielded from bacteria and other nasty stuff. However, like all good things, their lifespan is limited to about 100 days.

“It is estimated that living neurons use millions of times less energy than current digital processors.

The appeal of using living neurons is clear: they consume millions of times less energy than digital processors, which means they could be the key to making AI both smarter and greener. But as much as we love the idea of a brain-powered supercomputer, FinalSpark admits that bio-computers are still very much a black box, making programming a bit of a head-scratcher.

For those of you wondering if we’re one step closer to Skynet—don’t worry. We’re not there yet. But keep an eye on this space; we might just be witnessing the birth of a new kind of processor that’s less about ones and zeros and more about brainpower. Or, as FinalSpark might say, “Move over, silicon—neurons are taking the wheel.”