NextFin News - In a Melbourne laboratory, a cluster of several hundred thousand living human brain cells has successfully learned to navigate the demon-infested corridors of the 1993 video game Doom. The achievement, announced by Australian startup Cortical Labs on March 10, 2026, marks a significant leap from the group’s previous success with the simpler game Pong. By integrating biological neurons with silicon circuitry via their "CL1" platform, researchers have created a functional biocomputer that processes information not through binary logic gates, but through the adaptive, electrical signaling of living tissue.
The "DishBrain" system utilizes between 200,000 and 800,000 neurons derived from human induced pluripotent stem cells. These cells, originally sourced from blood donations, are grown onto a microelectrode array that both stimulates the culture and records its responses. To "play" the game, the neurons receive electrical pulses representing the game state—such as the location of an enemy—and their subsequent electrical spikes are translated into in-game movements. While the biological culture is not yet a match for a human player, Dr. Alon Loeffler, a scientist at Cortical Labs, confirmed that the cells outperform random chance, demonstrating a rudimentary ability to identify targets and adjust their behavior based on feedback.
This biological approach to computing addresses a looming crisis in the semiconductor industry: the staggering energy demands of artificial intelligence. A standard high-end AI chip can consume hundreds of watts and requires massive cooling infrastructure. In contrast, the CL1 biocomputer operates on approximately 30 watts, including its life-support systems. The neurons themselves are incredibly efficient, requiring only a nutrient-rich medium to maintain their "hardware." As global data centers are projected to consume more than 1,000 terawatt-hours of electricity annually by the end of the decade, the prospect of "wetware" that learns with a fraction of the energy footprint is no longer a fringe curiosity but a commercial necessity.
The implications extend far beyond gaming. Cortical Labs is already positioning the CL1 as a "code-deployable" biological computer for the pharmaceutical industry. Currently, drug development relies heavily on animal models that often fail to replicate human neurological responses, leading to a 90% failure rate in clinical trials for brain-related medications. By testing compounds directly on human neuronal cultures that are actively performing tasks, researchers can observe how a drug affects cognitive function in real-time. This could potentially eliminate the need for thousands of animal subjects while providing more accurate data for treating Alzheimer’s or epilepsy.
However, the rise of biocomputing brings a thicket of ethical and philosophical dilemmas. Critics argue that as these cultures grow in complexity, the line between a "biological processor" and a "sentient entity" begins to blur. While the current 800,000-cell cultures are far smaller than the 86 billion neurons in a human brain, the trajectory is clear. Cortical Labs has already announced plans to build biological data centers in Melbourne and Singapore, stacking these neuronal chips into racks. The question is no longer whether we can build computers from human cells, but whether we should grant these "mini-brains" a different moral status than the silicon chips they are designed to replace.
The commercial landscape is shifting as well. Cortical Labs has launched the "Cortical Cloud," allowing early-access users to deploy code directly to living neurons via the internet. This democratization of biocomputing suggests a future where software engineers might choose between a GPU and a biological culture based on the task at hand. For navigating unpredictable, changing environments—a task where traditional AI often struggles—the biological system’s innate plasticity offers a distinct advantage. The demons in Doom may be digital, but the intelligence hunting them is now undeniably, and perhaps uncomfortably, alive.
Explore more exclusive insights at nextfin.ai.
