A few years ago, scientists taught a bunch of neurons in a petri dish to play Pong. That was weird enough. But now, those neurons have leveled up. They’re not just playing games—they’re forming the foundation of the world’s first biological computer. Meet the CL1, a machine powered by living human brain cells, designed by Australian company Cortical Labs. Spotted at MWC 2025, the CL1 isn’t some futuristic thought experiment; it’s real, it’s commercialized, and it’s challenging everything we thought we knew about AI.
You’d expect something this advanced to be a towering, sci-fi supercomputer. It’s not. The CL1 is a sleek, self-contained unit that fuses human neurons with silicon hardware. Those neurons don’t just store data—they actively form connections, rewire themselves, and learn dynamically, just like the networks inside your brain. Traditional AI models, no matter how complex, rely on brute-force computation and massive datasets. The CL1 does something fundamentally different: it processes information biologically, adjusting in real time without needing thousands of GPUs churning through petabytes of data.
Designer: Cortical Labs
The secret lies in a technology called Synthetic Biological Intelligence (SBI). Think of it as a hybrid between human cognition and machine efficiency. Lab-grown neurons sit on an electrode array, where they’re fed electrical signals and respond by strengthening or weakening connections, much like how learning happens in a human brain. Unlike rigid AI models that require pre-programmed rules, the CL1’s neurons adapt organically. It’s like training a dog versus coding a robot—you don’t have to tell the neurons what to do explicitly. They figure it out.
This adaptability could have profound implications. In drug discovery, for example, researchers could test how neurons react to different compounds in real time, potentially speeding up treatments for diseases like Alzheimer’s and Parkinson’s. Unlike traditional cell cultures, which exist in static environments, these neural networks behave like actual brain tissue. Scientists wouldn’t just be observing chemical reactions; they’d be studying cognition in action. That level of biological realism could make clinical trials faster, cheaper, and more effective.
A close-up of the chip, which has actual human brain cells integrated into the silicon
This bio/tech hybrid forms an entirely new kind of artificial intelligence
There’s also a practical advantage: energy efficiency. AI as we know it is a power-hungry beast, with models like GPT-4 requiring warehouses of GPUs burning through electricity at an alarming rate. The CL1, by contrast, is a biological marvel. A full server rack of these units uses about the same energy as a high-end gaming PC. Human neurons are naturally optimized for low-power, high-efficiency processing—something silicon has never quite been able to replicate. This shift from artificial neural networks to actual biological ones could make AI development dramatically more sustainable.
But let’s be real—melding living brain cells with hardware raises all kinds of ethical and philosophical questions. Are these neurons “thinking”? Could they develop some form of awareness? So far, the answer is no. The networks inside the CL1 are highly specialized, built for processing information rather than conscious thought. But as the technology advances, those questions will only get louder. Cortical Labs has put ethical safeguards in place, ensuring that SBI development remains within controlled parameters. Still, the idea of neurons firing away inside a machine is enough to unsettle even the most enthusiastic tech optimists.
Beyond neuroscience and medicine, the CL1 could reshape robotics and autonomous systems. Traditional AI struggles with unpredictable environments because it lacks true adaptability. A biological AI system, on the other hand, could respond to its surroundings more fluidly, learning from direct interaction rather than static training data. Imagine a self-driving car that doesn’t just follow programmed rules but intuitively learns how to navigate new situations on the fly. That’s the kind of intelligence SBI could make possible.
For now, Cortical Labs is making the CL1 available through a cloud-based platform, allowing researchers to access SBI technology remotely. The hardware itself isn’t cheap—each unit costs around $35,000—but it’s already far more affordable than similar biotech research tools, some of which exceed $80,000. The long-term goal? Making SBI a mainstream technology, one that scientists, engineers, and innovators can build on just as easily as they do with traditional AI today.
This is the moment where AI stops being purely artificial. When neurons and circuits start working together, intelligence itself takes on a new form—one that’s more fluid, more adaptable, and perhaps even closer to what we call “thinking.” Where this leads, no one knows yet. But one thing is certain: the future of AI just got a lot more organic.
The post World’s First “Biological Computer” Runs AI Models On Actual Human Brain Cells first appeared on Yanko Design.