Singapore is putting living neurons into the data center conversation. Digital infrastructure company DayOne has launched what it describes as Singapore’s first Biological Data Center Prototype with Cortical Labs and the Yong Loo Lin School of Medicine at the National University of Singapore, exploring whether biological computing could complement silicon infrastructure as AI workloads push energy and capacity requirements higher.

The next phase of AI infrastructure may not be built entirely from silicon.

DayOne Data Centers, a Singapore-headquartered digital infrastructure company, has launched a Biological Data Center Prototype in partnership with Cortical Labs and the Yong Loo Lin School of Medicine, National University of Singapore (NUS Medicine). The project combines living human neurons with conventional computing hardware to investigate a radically different approach to processing information.

The initiative comes as Singapore expands its AI and data center capacity while facing a familiar constraint: electricity and environmental impact.

The prototype deploys 20 Cortical Labs CL1 biological computing units in an independently operated server rack hosted within infrastructure supported by DayOne. According to the companies, it is the world’s first independently operated biologically integrated server rack.

Rather than replacing conventional servers, the project is testing whether biological computing can eventually serve as a specialized complement to silicon-based infrastructure.

That distinction matters.

Biological computing, sometimes referred to as wetware computing, uses living neurons grown from stem cells and interfaced with electronic systems. The neurons receive electrical stimulation, respond to signals and adapt based on their interactions. Cortical Labs’ approach combines these biological networks with software and silicon electronics to create computing systems capable of learning from their environment.

The concept is still highly experimental. But its potential appeal is straightforward: biological systems process information using mechanisms developed through evolution rather than conventional transistor-based architectures, potentially allowing some workloads to be handled with substantially lower energy consumption.

For AI infrastructure operators, energy efficiency has become a strategic issue rather than simply an environmental one.

The International Energy Agency estimates that electricity consumption from data centers worldwide could more than double by 2030, reaching roughly 945 terawatt-hours. AI is one of the major drivers behind that growth, particularly as accelerated computing and high-performance AI systems require increasingly large amounts of electricity.

That creates an opening for alternative computing architectures.

DayOne’s prototype is designed to explore applications including neuro-inspired AI, biomedical modeling, drug discovery and neurological disease research. Cortical Labs also identifies areas such as robotics, cybersecurity and fraud detection as potential future applications.

The immediate value, however, may lie as much in research as in data center economics.

NUS Medicine brings neuroscience and biomedical expertise to the project, while Cortical Labs supplies the biological computing platform. DayOne contributes the infrastructure required to operate the system within a live environment rather than an isolated laboratory setup.

That combination creates a bridge between experimental neuroscience and digital infrastructure.

Professor Rickie Patani, a professor of neuroscience at NUS Medicine and director of the Neurobiology Programme at the NUS Life Sciences Institute, described the collaboration as an opportunity to study learning and adaptation at their biological source while investigating applications in drug discovery and neurological disease research.

The biomedical angle could prove particularly important.

Conventional AI models are powerful but can require substantial quantities of training data and computing resources. Biological neural networks, by contrast, could offer researchers a way to investigate learning and adaptation using living cells themselves. The potential advantage is not that neurons will replace GPUs for general-purpose AI, but that they may be useful for specific problems where biological adaptation is valuable.

This makes the project conceptually different from the GPU-centric infrastructure being built by companies such as NVIDIA, Microsoft, Amazon Web Services and Google.

Those platforms are scaling AI primarily through increasingly powerful accelerators, networking, memory and data center architectures. Biological computing represents a different direction: changing the underlying computational substrate rather than simply increasing the performance of silicon.

For enterprises, that distinction is important. Biological computing is unlikely to become a drop-in replacement for existing cloud infrastructure in the near term. Its more realistic trajectory is as a specialized computing architecture for selected workloads.

The potential applications could include systems that need to learn continuously from limited information, simulate biological processes or operate efficiently under constrained power conditions.

Cortical Labs founder and CEO Hon Weng Chong said the company’s objective is to identify situations where biological systems can learn from less data and adapt as conditions change.

Whether those advantages translate into commercially meaningful performance remains an open question.

There are also significant operational challenges. Biological computing requires maintaining living cells, managing biological variability and developing new standards for reliability, reproducibility, security and system lifecycle management. Data center operators accustomed to predictable silicon hardware will need entirely different operational models if wetware systems move beyond research environments.

Singapore provides an unusually relevant testbed for that transition.

The country is one of Asia’s major data center markets but has also imposed constraints designed to improve energy efficiency and manage the environmental impact of digital infrastructure. Its Green Data Centre Roadmap aims to support additional data center capacity while improving energy efficiency and sustainability.

That policy environment makes alternative computing architectures strategically interesting.

DayOne says the biological computing initiative forms part of its wider Singapore investment strategy. Its first local data center, SG1, reached structural topping-out in May 2026 and is targeting operations in early 2027. The company is also developing a Global Operations Command Center in Singapore.

DayOne says it has secured approximately 2.1 gigawatts of bookings across its nine-market Asia-Pacific and European platform. That scale gives the company an established position in conventional digital infrastructure while it experiments with a very different form of compute.

The larger question is whether biological computing can move from an intriguing scientific demonstration into an economically viable infrastructure category.

For now, Singapore’s prototype is better viewed as an experiment than an alternative to the GPU data center. But as AI pushes conventional infrastructure toward limits involving power, cooling, land and grid capacity, experiments with entirely new computational substrates are becoming harder for the industry to ignore.

The future of AI infrastructure may not be silicon versus biology. It may be a more heterogeneous computing landscape in which each architecture handles the workloads it is best suited to solve.

Market Landscape

AI infrastructure is entering a period of architectural experimentation.

The dominant model remains silicon-based accelerated computing, led by GPUs and specialized AI accelerators from companies such as NVIDIA and competing infrastructure providers. Hyperscalers including Microsoft, Amazon and Google are simultaneously investing in custom silicon, data center optimization and renewable-energy strategies.

Biological computing represents a much earlier-stage alternative.

Its potential advantages include:

• Energy efficiency: Certain biological processes may perform specialized computations at substantially lower power than conventional digital systems.
• Adaptive learning: Neural cultures can respond and adapt to changing stimuli, creating potential applications in adaptive AI.
• Biomedical research: The same systems used for computation can provide models for studying neurological processes and drug responses.
• Specialized AI: Wetware may eventually complement GPUs for workloads involving sparse data, adaptation or biological simulation.
• Infrastructure diversification: New architectures could reduce dependence on continual increases in conventional compute density.

The market remains experimental, however. Biological computing must demonstrate consistent performance, scalability, reliability and economic viability before it can compete meaningfully with established AI infrastructure.

The broader trend is clear: AI infrastructure is becoming heterogeneous. GPUs, CPUs, custom accelerators, neuromorphic systems, edge processors and potentially biological computing could ultimately coexist rather than compete for a single dominant architecture.

Top Insights

• DayOne’s Singapore prototype combines Cortical Labs’ biological computing with data center infrastructure, testing whether living neurons can complement silicon AI systems.
• The project targets energy-efficient computing as AI data centers face growing electricity demands, particularly for specialized workloads requiring adaptation and limited training data.
• NUS Medicine adds neuroscience expertise, creating potential applications spanning drug discovery, neurological research, biomedical modeling and neuro-inspired artificial intelligence.
• Biological computing remains experimental, meaning enterprises should view wetware as a specialized research opportunity rather than a near-term replacement for GPUs or cloud infrastructure.
• Singapore’s sustainability policies make it a strategic environment for testing alternative computing architectures as governments balance AI expansion with energy constraints.

Power Tomorrow’s Intelligence — Build It with TechEdgeAI