Melbourne, Australia – A unique datacenter is challenging conventional computing norms, not with silicon, but with living brain cells. Cortical Labs, a biotechnology company based in Melbourne, is pioneering biological computing, and the daily routine at its facility involves a process unlike any other: replenishing computers with a fluid mimicking cerebrospinal fluid. This isn’t science fiction; it’s the reality of a nascent industry aiming to revolutionize processing power and energy efficiency.
The need for this specialized fluid stems from the very nature of the “computers” themselves – lab-grown neurons. According to Hon Weng Chong, CEO and founder of Cortical Labs, the neurons deplete oxygen and glucose from the liquid, necessitating a daily top-up. Beyond the fluid, technicians also carefully adjust the atmospheric composition around the biological computers, maintaining approximately five percent oxygen through a mixture of nitrogen and carbon dioxide – conditions optimized for neuronal activity.
While still in its early stages, biological computing holds the promise of surpassing traditional computers in certain areas. Chong asserts that these neuron-based systems can learn and adapt within simulated environments more rapidly than classical computers, generating novel solutions rather than simply reordering existing information like large language models (LLMs). The potential for significantly reduced energy consumption is a key driver behind this innovative approach.
However, scaling biological computing faces significant hurdles. A major bottleneck is the limited availability of facilities capable of providing and handling the necessary cell cultures. Chong highlights the need for a “cell foundry” – an equivalent to TSMC for biological components – to make this technology more accessible. To bridge this gap, Cortical Labs has launched a cloud service, offering access to its biological computing capabilities.
The cloud platform, built on 120 CL1 units, allows users to run code – whether through Jupyter Notebooks or Python scripts – on biological hardware. Access is granted via credit card payment, but unlike hyperscale cloud providers, preparation for each job takes approximately a week. This involves sourcing the specific cell type requested by the customer and establishing the optimal physical environment. Most users, Chong anticipates, will initially rent a little cluster of three or four CL1 units to replicate results and maintain experimental control.
Early adopters are expected to be research labs lacking their own CL1 infrastructure, or organizations exploring unconventional computing solutions. Chong points to the growing interest in quantum computing as a parallel, with some organizations beginning to investigate biological computing as a potential future technology. An article in Information Age noted that Cortical Labs has been building up its CL1 infrastructure since September 2025.
From Pong to DOOM: Demonstrating Biological Learning
Cortical Labs has already demonstrated the learning capabilities of its biological computers. The company gained attention in 2022 with “DishBrain,” a system where neurons learned to play Pong. More recently, the team showcased their machines’ ability to play DOOM, building upon research detailed in a 2022 paper titled “In vitro neurons learn and exhibit sentience when embodied in a simulated game-world.”
The research involved placing a biological neural network (BNN), composed of both human and rodent stem cells, on high-density multielectrode arrays. As the paper explains, this setup allows for communication between silicon hardware and the BNN through electrical stimulation and recording, leveraging the neurons’ inherent ability to share information. This foundational work led to the development of the commercially available CL1 biological computer.
The Challenges of Biological Hardware
Utilizing a CL1 requires careful consideration of cell selection, choosing lines with genetic traits suited to the specific computational task. Beyond that, the ongoing need for specialized fluids and gas mixtures adds complexity. While automation to streamline these processes is a future goal, Chong believes users will currently accept the hands-on approach. He also admits to a degree of caution regarding granting complete autonomy to biological computers, joking about being uncomfortable with them controlling their own destiny.
Cortical Labs, founded in 2019 by Hon Weng Chong, has secured funding including a $1.62 million seed round in June 2019 and a $10 million round led by Horizons Ventures in April 2023, according to Wikipedia. The company’s team also includes Chief Scientific Officer Brett Kagan and Chief Technology Officer David Hogan.
The development of biological computing is still in its infancy, but Cortical Labs is actively working to overcome the existing hurdles and unlock the potential of this groundbreaking technology. The company’s cloud service represents a significant step towards wider accessibility, allowing researchers and organizations to explore the unique capabilities of neuron-based computation. As the field matures, the demand for specialized cell foundries and automated systems will likely increase, paving the way for a new era of computing.
The future of computing may very well be alive, and Cortical Labs is at the forefront of this exciting and rapidly evolving field. Share your thoughts on the potential of biological computing in the comments below.
