Human Brain Cells Now Power Data Centers in Revolutionary Computing Breakthrough
Europe, Saturday, 4 April 2026.
Australian startup Cortical Labs has achieved a remarkable milestone by deploying 200,000 living human neurons in operational data centers across Melbourne and Singapore. Their CL1 system transforms months of specialized laboratory work into mere hours, combining lab-grown brain cells with silicon chips to create biological processors that cost $35,000 per unit. These living neurons have demonstrated extraordinary capabilities, learning to play complex video games like Doom within a week while consuming dramatically less energy than traditional computers. The breakthrough represents a fundamental shift from conventional silicon-based computing to ‘wetware’ that leverages the natural processing power and adaptability of human brain cells for real-world applications.
The Technology Behind Living Processors
The CL1 system operates by growing neurons from human stem cells and placing them on specialized chips equipped with microelectrodes that send and receive electrical signals [1]. These 200,000 living human neurons are maintained in a precisely controlled environment within sealed chambers, where life-support systems monitor critical conditions: 36.8°C temperature, 4.7% carbon dioxide, and 5.8% oxygen levels [2]. The process begins with human adult cells that are reversed to an embryonic stem cell state, then grown into neurons and plated on a microelectrode array featuring 59 electrodes [2]. According to Brett J. Kagan, chief scientific officer and chief operating officer at Cortical Labs, “All you need is a little bit of blood or some skin, and you can generate an indefinite supply of these cells that you can then turn into neurons” [1].
Operational Scale and Current Deployment
Cortical Labs has successfully deployed 120 units of their CL1 system to run a small data center in Melbourne, Australia [1]. The company has also established biological computing facilities in both Melbourne and Singapore, offering remote access to multiple CL1 units through their Cortical Cloud platform [1][2]. Each CL1 instance consists of 200,000 living human neurons interfaced with a silicon chip, with current pricing set at $35,000 per biological computer unit [2][3]. The platform represents a significant advancement from traditional laboratory work, as Kagan explains: “We’re using these cells more like an engineering approach to build something that’s never really existed before and might have properties that we’ve never been able to use before. And so far, the results are very exciting” [1].
Extraordinary Learning Capabilities
The living neurons have demonstrated remarkable learning abilities that distinguish them from conventional silicon-based systems. In documented experiments, 200,000 human neurons grown on a chip learned to navigate the complex video game Doom in under a week [5]. Even more impressive, earlier experiments showed these neurons learning to play Pong within minutes, demonstrating real-time learning capabilities [4]. The breakthrough came through innovative training methods developed by Cortical Labs, which used sine waves as rewards and white noise as punishment signals, based on Karl Friston’s Free Energy Principle [2]. Independent developer Sean Cole created an interface between Doom and living neurons in under a week using Python, showcasing the accessibility of the technology [2].
Energy Efficiency and Competitive Advantages
The biological approach offers significant energy advantages over traditional computing systems. Hon Weng Chong, CEO of Cortical Labs, emphasizes that “Biology has to obey the laws of physics — you cannot consume more energy than what you can hunt or forage. And we know what happens when you do that. You die” [2]. To illustrate this efficiency, biological systems demonstrate extraordinary performance ratios: a dragonfly with one million neurons can predict prey interception points in a tenth of a second using only 0.5 milliwatts of energy [2]. Kagan notes that “Biology is incredibly energy efficient. We [humans] don’t require huge amounts of data” [1]. The CL1 system differs fundamentally from conventional silicon chips by using living cell cultures, termed “wetware,” which allows for direct interaction with neurons and provides fluid rather than crystallized intelligence [1][2].