{"slug": "national-university-of-singapore-launches-world-first-data-center-powered-by", "title": "National University of Singapore launches world-first data center powered by human brain cells", "summary": "The National University of Singapore's Life Sciences Institute launched the world's first biologically integrated server rack on July 16, using 20 Cortical Labs CL1 units powered by lab-grown human neurons that consume about 30 watts each, compared to over 700 watts for an Nvidia H100 GPU. Built with NUS Medicine's Yong Loo Lin School of Medicine, Melbourne-based Cortical Labs, and Singapore data center operator DayOne, the system targets energy-efficient AI processing and biomedical research, with DayOne considering scaling to 1,000 units after raising $4.5 billion in June at a $20 billion valuation.", "body_md": "Via educatly.com\n\n# National University of Singapore launches world-first data center powered by human brain cells\n\nThe 20-unit biological computing system uses lab-grown neurons that consume a fraction of the energy of traditional servers, marking a new frontier in sustainable AI infrastructure.\n\nA data center in Singapore is now running on human brain cells. That sentence reads like the opening crawl of a sci-fi movie, but it’s a real thing happening at the National University of Singapore’s Life Sciences Institute, where 20 biological computing units containing lab-grown human neurons went live on July 16.\n\nThe system, publicly unveiled on August 6, represents the world’s first independently operated biologically integrated server rack. It was built through a collaboration between NUS Medicine’s Yong Loo Lin School of Medicine, Melbourne-based biotech startup Cortical Labs, and Singapore data center operator DayOne.\n\n## How it works, without the sci-fi jargon\n\nEach of Cortical Labs’ CL1 units houses between 200,000 and 800,000 human cortical neurons grown in a lab, not harvested from anyone’s skull. These neurons sit on silicon-based microelectrode arrays, creating a hybrid system where biological tissue and traditional chips work together.\n\nThe cells need feeding every three days and are kept alive by dedicated life-support equipment that maintains precise gas mixtures of CO2, oxygen, and nitrogen, along with carefully controlled temperatures. Their operational lifespan runs about six months before they need to be replaced.\n\nThe system uses no consciousness-generating architecture. These are clusters of neurons performing computational tasks, not thinking or experiencing anything.\n\n## The energy math is striking\n\nEach CL1 unit consumes roughly 30 watts, life support included. A high-end Nvidia H100 GPU under full load demands upwards of 700 watts. That’s more than a 20x difference in power consumption per unit.\n\nThe 20-unit prototype is small by any data center standard. But DayOne, which raised $4.5 billion in June and achieved a $20 billion valuation, is reportedly considering scaling the deployment to 1,000 units. At 30 watts per unit, that entire biological computing cluster would consume 30 kilowatts.\n\n## From Pong to production\n\nCortical Labs previously demonstrated that its biological systems could learn to play basic video games, a proof of concept that showed neurons could adapt to digital inputs and improve their responses over time.\n\nThe NUS deployment marks the first time Cortical Labs has set up a biological computing facility outside its home base in Melbourne. The partnership was announced in March, and the roughly four-month timeline from announcement to operational prototype suggests the underlying technology has matured past the purely experimental phase.\n\nThe system targets two broad categories of use: energy-efficient AI processing and biomedical research, including disease modeling and drug screening. Having living human neurons in a controlled computational environment creates a unique platform for testing how those neurons respond to pharmaceutical compounds or disease conditions.\n\n## What this means for AI infrastructure\n\nDayOne’s ambition to scale from 20 to 1,000 units would represent a 50x expansion. Regulatory approval remains a key variable, as deploying living human tissue in commercial computing environments raises questions that existing frameworks weren’t designed to address.\n\nThe six-month neuron lifespan introduces an operational model that looks nothing like traditional infrastructure. Server hardware might run for years before replacement. Biological computing units would need regular cell replenishment, creating ongoing costs and logistics that don’t exist in conventional data centers.\n\n**Disclosure:** This article was edited by Editorial Team. For more information on how we create and review content, see our\n\n[Editorial Policy](https://cryptobriefing.com/editorial-policy/).", "url": "https://wpnews.pro/news/national-university-of-singapore-launches-world-first-data-center-powered-by", "canonical_source": "https://cryptobriefing.com/singapore-brain-cell-data-center/", "published_at": "2026-08-24 11:21:50+00:00", "updated_at": "2026-08-24 11:43:24.090191+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-infrastructure", "ai-research"], "entities": ["National University of Singapore", "Life Sciences Institute", "NUS Medicine's Yong Loo Lin School of Medicine", "Cortical Labs", "DayOne", "Nvidia H100"], "alternates": {"html": "https://wpnews.pro/news/national-university-of-singapore-launches-world-first-data-center-powered-by", "markdown": "https://wpnews.pro/news/national-university-of-singapore-launches-world-first-data-center-powered-by.md", "text": "https://wpnews.pro/news/national-university-of-singapore-launches-world-first-data-center-powered-by.txt", "jsonld": "https://wpnews.pro/news/national-university-of-singapore-launches-world-first-data-center-powered-by.jsonld"}}