Your Next Computer Might Be Alive — And Nobody's Ready

A computer that runs on living brain cells sounds like a Black Mirror episode. It isn’t. Australian startup Cortical Labs began selling the CL1 last year — a $35,000 machine with roughly 800,000 living human neurons growing on a silicon chip. The technology is shipping. The ethics aren’t even drafted.

What “Wetware” Actually Means

Conventional AI mimics neurons digitally. Wetware uses the real thing — neurons differentiated from human stem cells, cultured on electrode-laced chips, fed and queried with electrical signals.

The pitch is energy. A human brain runs on roughly 20 watts. Training a frontier LLM burns megawatts. That’s a million-fold gap, and it’s the kind of efficiency curve that makes investors lose sleep.

Researchers have already taught cultured neuron clusters to play Pong, navigate mazes, and recognize patterns. In some narrow tasks, they learn faster than digital networks. The catch: we cannot tell whether the cells feel anything while doing it.

Consciousness Is the Question Nobody Can Answer

800,000 neurons is roughly insect-tier — trivial next to the human brain’s 86 billion. Plenty of researchers shrug and say there’s nothing there to worry about.

The opposing camp points to a basic gap in neuroscience: we still don’t know at what threshold subjective experience emerges. The “hard problem” of consciousness remains unsolved. If a cultured cluster experiences even a faint flicker of awareness or distress, we have no instrument to detect it.

Is wiping a neural culture morally equivalent to clearing a hard drive? Is using electrical “punishment” as a training signal acceptable? The field has no consensus, and the machines are already on sale.

Whose Cells Are These, Anyway?

Most wetware uses induced pluripotent stem cells (iPSCs) — typically reprogrammed from a donor’s skin sample, then differentiated into neurons. The standard medical research consent form does not say “your cells may end up as compute substrate in a commercial product.”

It gets weirder. If a system built from someone’s neurons eventually exhibits patterns that echo their cognition, who owns that? US biospecimen law — already shaky since Henrietta Lacks and Moore v. Regents — has no framework for this. Neither does the GDPR.

The Regulatory Vacuum

The EU AI Act regulates digital systems. US executive orders on AI do the same. Neither addresses computing on living neural tissue in any direct way. The FDA touches organoids only in a research context. There is no agency clearly on the hook.

Meanwhile Cortical Labs is preparing cloud access to wetware compute. Switzerland’s FinalSpark already offers remote sessions on biological neural networks to academic clients. The product exists. The rulebook doesn’t.

Researchers at Johns Hopkins and Yale have floated an “organoid rights” framework — a precautionary principle argument: if the probability of sentience isn’t zero, treat it as if it might be nonzero. Industry pushes back hard, calling it overreach for tissue that’s “less complex than a fruit fly.”

What Has to Happen Before the Market Locks In

Technology routinely outruns ethics — IVF, CRISPR, brain-computer interfaces all followed this script. Wetware is different in one important way: once living neural tissue becomes commercial infrastructure, pulling it back becomes politically and economically near-impossible. Jobs, supply chains, and IP portfolios calcify around it.

The useful debate isn’t ban-versus-allow. It’s whether we get consciousness assessment standards, donor consent protocols, disposal guidelines, and liability frameworks in place before the industry hardens around the current free-for-all.

The honest question isn’t whether you’d buy a computer that might, on some level, be alive. It’s whether that decision will still be yours to make a few years from now — or whether the market will have made it for all of us.