Human Brain Cells: AI’s Unexpected Power Source? | Doom-Playing Neurons & Bio-Computers

The relentless demand for computing power, driven largely by the expansion of artificial intelligence, is pushing researchers and companies to explore radical new approaches to data center design. Among the most unconventional: leveraging the biological processing power of human neurons. Australia-based Cortical Labs is leading the charge, announcing plans to build two “biological” data centers – one in Melbourne and another in Singapore – powered by chips embedded with 200,000 living human brain cells. This development, initially highlighted by the company’s demonstration of a neuron-filled chip successfully playing the video game Doom, represents a significant step toward a potentially more energy-efficient computing future.

How Neuron-Based Computing Works

Traditional computers rely on silicon transistors to process information as binary code (0s and 1s). Cortical Labs’ approach, however, utilizes the inherent parallel processing capabilities of biological neurons. The company grows human neurons – derived from adult stem cells harvested from blood samples – directly on microchips equipped with multi-electrode arrays. These arrays stimulate the neurons and record their electrical activity, effectively translating neuronal responses into computational output. As CEO Hon Weng Chong explained to Bloomberg, a single CL1 computer consumes less power than a handheld calculator. This efficiency stems from the neurons’ ability to learn and adapt with minimal energy expenditure, a stark contrast to the power-hungry demands of conventional AI chips. The process involves converting blood cells into stem cells, then differentiating those stem cells into neurons, which are then cultivated directly on the microchip surface. The New York Post details this process, framing it as a modern-day Frankensteinian endeavor.

Beyond Gaming: Scaling Biological Computation

Whereas the initial demonstration of neurons playing Doom captured public attention, the long-term goal extends far beyond entertainment. Cortical Labs envisions these biological chips powering a new generation of AI data centers. The Melbourne facility is planned to house approximately 120 CL1 units, while the Singapore data center, a collaboration with the National University of Singapore, will initially feature 20 units, with ambitions to scale to 1,000. This expansion aims to make Cortical Labs’ brain-computing service more widely accessible, as noted by Michael Barros at the University of Essex, UK, who already utilizes the company’s cloud services for his research. New Scientist reports that this accessibility is a key differentiator, as building and utilizing biological computers is often challenging for individual research groups.

The Energy Imperative and Data Center Demand

The push for neuron-based computing is fueled by the escalating energy consumption of traditional data centers. These facilities, essential for powering cloud services and AI applications, require vast amounts of electricity, often straining local power grids and driving up costs. The announcement from Cortical Labs comes at a time when tech giants like Google, Amazon, Microsoft and OpenAI are investing billions in new data center infrastructure to support their AI models. Gizmodo highlights this context, noting that even President Trump has previously intervened to secure commitments from tech companies regarding power costs associated with their data center projects.

Limitations and Ethical Considerations

Despite the potential benefits, significant hurdles remain. The number of neurons used in these chips – 200,000 – pales in comparison to the estimated 60 to 99 billion neurons found in the human brain. This difference in scale limits the complexity of tasks these bio-computers can currently handle. Maintaining a living system introduces unique challenges. The Singapore facility, for example, requires a dedicated system for replenishing cerebrospinal fluid, as reported by The Register.

Ethical considerations also loom large. The sourcing of neurons from human stem cells raises questions about consent and potential exploitation. While Cortical Labs states that the cells are derived from adult volunteers, the long-term implications of utilizing human biological material for computational purposes require careful scrutiny. The potential for unforeseen consequences, as well as the broader societal impact of blurring the lines between biology and technology, necessitate ongoing dialogue and responsible development.

What Comes Next: Scaling, Regulation, and Research

The immediate next steps for Cortical Labs involve securing regulatory approval for the expansion of the Singapore data center and refining the neuron cultivation process to improve scalability and reliability. Further research is needed to optimize the communication between neurons and the microchip arrays, as well as to explore the potential for integrating these biological chips with traditional computing systems. The company’s success will likely depend on its ability to demonstrate a clear performance advantage over conventional computing methods, particularly in specific applications where the parallel processing capabilities of neurons can be fully leveraged. The broader field of biological computing is still in its early stages, and continued investment in fundamental research will be crucial for unlocking its full potential. The company is also focused on expanding its cloud-based brain-computing service, making the technology accessible to a wider range of researchers and developers.