Featured Image: yourNEWS Media Illustration
A Singapore prototype combines lab-grown human neurons with conventional electronics to process information, offering an early look at biological computing while raising unresolved questions about scalability, energy use and ethics.
By yourNEWS Media Newsroom.
Researchers are moving an unusual form of computing out of the laboratory and toward something resembling a data center: machines that use living human neurons as part of the information-processing system.
A prototype unveiled in Singapore last month has been described by its developers as a “biological data center”, combining living neural tissue with conventional silicon electronics.
Instead of relying entirely on processors made from semiconductors, the system incorporates clusters of human neurons grown in laboratory conditions and connected to electronic interfaces capable of stimulating the cells and recording their responses.
The technology is often called “wetware.”
According to a report by The Defender examining the project, the living neurons are derived from human stem cells and maintained under controlled laboratory conditions while communicating with conventional computer hardware.
The cells themselves originate from reprogrammed donor skin or blood cells. Once converted into stem cells, they can be differentiated into neurons and grown directly on specialized computing hardware.
The neural cultures reportedly remain viable for about six months.
During that period, life-support equipment keeps them at controlled temperatures, supplies nutrients and maintains sterile conditions necessary for the cells to survive.
Arrays of electrodes stimulate the neurons and record their electrical responses. Software then interprets the resulting activity and converts it into information conventional computers can use as inputs and outputs.
The National University of Singapore’s Yong Loo Lin School of Medicine developed the prototype with Cortical Labs and data-center company DayOne.
The university described the project as “a powerful glimpse into how biology and technology can come together to transform the future of computing.”
Not everyone views the development so enthusiastically.
Patrick Wood, editor-in-chief of Technocracy News, called the prototype “creepy.”
Neurons Become Part of the Computing Hardware
The central idea behind the technology is that living neurons process information differently from conventional chips.
Traditional processors execute programmed operations through fixed electronic circuitry. Biological neurons form connections, respond to stimulation and alter their signaling patterns over time.
Cortical Labs argues that those properties could eventually make neural computing useful for problems involving learning and adaptation.
Each of the company’s CL1 units contains approximately 200,000 human neurons grown on a silicon chip equipped with a microelectrode array.
The electrodes send electrical signals into the neural culture and record how the neurons respond. Software interprets those changing patterns and creates an interface between the biological tissue and conventional computers.
Cortical Labs previously demonstrated versions of its technology by training neural cultures to interact with simple video-game environments, including Pong and Doom.
Those experiments were intended to show that neurons growing outside a human body can respond dynamically to electronic feedback rather than merely producing random electrical activity.
The company now appears to be taking that concept toward larger computing installations.
Developers contend that biological processors could eventually address some of the resource problems associated with rapidly expanding digital infrastructure and artificial intelligence.
The Energy Argument Remains Unproven at Scale
One of the most frequently cited advantages of biological computing is energy efficiency.
Today’s AI systems depend on enormous numbers of conventional processors operating inside facilities that can require large quantities of electricity and water.
By comparison, the human brain performs extraordinarily complex information processing while consuming roughly 20 watts of power.
That difference has encouraged researchers to ask whether living neurons could perform certain computational tasks using far less electrical energy than silicon-based AI systems.
But the comparison is incomplete.
The roughly 20-watt figure refers to energy consumed by the brain itself. A laboratory neural-computing system also requires pumps, temperature control, nutrient delivery, sterile environments, monitoring equipment, electronics and other infrastructure needed to keep the cells alive.
Publicly available claims about the prototype have not established how much energy the entire life-support system consumes.
That distinction is important because other technologies involving living laboratory-grown tissue have sometimes required considerable energy and supporting infrastructure. Critics have pointed to cultivated meat production as an example of a biological technology whose full resource requirements can be substantially greater than the cells themselves might suggest.
Another challenge is the pace at which conventional computing evolves.
Data centers can take years to design and construct, while AI hardware is advancing rapidly enough that some industry observers warn equipment can approach obsolescence before facilities are completed.
Biological computing would have to compete in that same environment.
Current Systems Remain Tiny Compared With Conventional Data Centers
Despite the “data center” label, the technology remains far removed from the scale of the facilities operated by companies such as Amazon, Microsoft, Google or Meta.
The present systems are “bench-scale.”
Cortical Labs has established a facility in Melbourne, Australia, and plans a larger installation in Singapore, but the Melbourne operation contains only about 120 units.
A conventional large-scale data center can contain tens of thousands of servers.
With approximately 200,000 neurons in each CL1 unit, even hundreds of biological processors would still represent a small experimental installation rather than a direct substitute for modern computing infrastructure.
That creates several technical problems.
Neurons have to remain alive and healthy. Biological cultures change over time. Individual units may not behave identically. The cells eventually die and must be replaced.
Reliability standards are also much higher for industrial computing than for laboratory demonstrations.
“Scaling living cells into something reliable enough for infrastructure use is a major engineering hurdle,” according to reporting on the project.
Claire Robinson, founder of GMWatch, said any attempt to expand the system substantially “will require huge numbers of cells, which will have to be regularly replaced, as they don’t last forever.”
The ability of cultured neurons to interact with Pong, Doom or other simple environments demonstrates that biological cells can participate in computational feedback loops, but it does not establish that they can perform the workloads handled by today’s commercial processors.
The gap between proof of concept and economically useful infrastructure remains substantial.
Cortical Labs Began With Medical Research Applications
Cortical Labs was founded in 2019 and had raised approximately $11 million as of March.
Its investors include Horizons Ventures, Blackbird Ventures, LifeX Ventures, Radar Ventures and In-Q-Tel, the CIA-funded venture capital organization.
The data-center concept represents an expansion of how Cortical Labs originally presented its technology.
When the CL1 system debuted in 2025, the company described it primarily as a “platform technology” for medical research, including drug discovery and disease modeling.
Researchers envisioned using laboratory-grown neural systems to study disorders such as epilepsy and Alzheimer’s disease and potentially test how brain cells respond to new therapies.
Turning the same technology toward general computing raises a different set of technical and ethical questions.
Living Computers Renew Debate Over Biological Technology
The concept of integrating living neural cells with machines fits within a much longer history of government, academic and commercial research into the brain.
Wood said the work reminded him of former President Barack Obama’s BRAIN Initiative, a major neuroscience program supported by the National Institutes of Health and agencies including DARPA.
The initiative sought to develop new technologies for mapping and understanding brain activity and was promoted by officials including former NIH Director Francis Collins.
A federal video explaining the initiative described some of those early goals in “On The Clock: The BRAIN Initiative”.
Wood criticized the direction of such research.
“The people have too much time on their hands, and if it wasn’t for public tax money, they would starve to death,” Wood said.
Robinson raised a broader concern about the boundaries between biological organisms and machines.
She said the project raises questions “about the wider ‘transhumanist’ landscape in which scientists and billionaire investors are trying to blur the line between living beings — including humans — and AI/computers.”
She added:
“We’ve all seen articles enthusing about AI robot soldiers — basically computers with a human appearance and attributes. There are worries about who is controlling these technologies, what their motivation is, what happens when something goes wrong and/or the robot gets out of control, and who is responsible and accountable for any harm caused.”
Robinson also argued that greater scrutiny should accompany development of the technology.
“We urgently need to demand transparency and accountability for where research and investment is being directed and engage in the ethical issues involved. We need to govern the technology before it governs us.”
Those concerns represent critics’ views rather than evidence that the current neural cultures possess consciousness or other human-like mental properties.
The CL1 systems use relatively small cultures of laboratory-grown neurons connected to electrodes. They are not miniature human brains and do not reproduce the anatomical organization or complexity of a human brain.
That distinction does not eliminate ethical questions, particularly if future systems become larger, more complex or more capable of adaptive behavior.
It does, however, place the current prototype in perspective.
The “biological data center” is presently an experimental computing system consisting of living neural cultures linked to conventional electronics, not a replacement for today’s large-scale cloud or AI facilities.
Its most important achievement may therefore be demonstrating that living human neurons can be engineered into repeatable computational hardware.
Whether that approach can become cheaper, more reliable or more energy-efficient than silicon remains unanswered.
Researchers will still have to demonstrate that the cells can survive predictably, perform useful workloads, scale far beyond hundreds of thousands of neurons per unit and deliver enough computational advantage to justify the equipment required to sustain living tissue.
Until then, the project represents less a finished alternative to conventional computing than an early test of an unusual proposition: that some of the processors inside future computers may be alive.
Original article
Comments