Cortical Labs builds brain cell data centers to tackle AI energy consumption crisis

Startup is building the first data centre to use human brain cells

Cortical Labs is building two data centres that will house its neuron-filled chips. The technology is still in the very early stages of development Data centres use huge amounts of energy and chips are in high-demand - could brain cells be the answer? Australia-based startup Cortical Labs has announced it is building two "biological" data centres in Melbourne and Singapore, stacked with the same neuron-filled chips it has demonstrated can play Pong or Doom. Cortical Labs is one of a few companies developing biological computers, consisting of neuronal cells wired up to microelectrode arrays that can stimulate and measure the response of cells when fed data. Earlier this month, the company demonstrated that its flagship computer, the CL1, could learn to play the game Doom in a week. Now, Cortical Labs has revealed two data centres that the company plans to build. The first in Melbourne will contain around 120 CL1 units, while the second, which is being built together with the National University of Singapore, will house 20 CL1s to start, which the company hopes will eventually contain 1000 units in a larger data centre, after regulatory approval. The company says this will allow them to expand their cloud-based brain computing service. Biological computers like the CL1 are being built and tested by research groups around the world, but they are often hard to build and not easy for others to use, says Michael Barros at the University of Essex. "We spend a lot of money and sweat to build these [systems]." "What [Cortical Labs] is doing is essentially allowing its biocomputer to be accessible at a large scale," says Barros, who already uses Cortical Labs' cloud services as part of his research. "They'll be the first ones to do that." Though these systems can be trained for relatively simple tasks, like playing Doom, the exact way in which these neurons function and how best to train them to perform tasks, like machine learning, is still unclear, says Reinhold Scherer, also at the University of Essex. "Having access to this allows you to explore learning, training and programming," says Scherer. "You don't program neurons like standard computers." Cortical Labs argue that its data centres will also require far less energy than typical computing systems, with each CL1 requiring around 30W, the company claims, rather than the thousands of watts than a state-of-the-art conventional AI chip. "When we scale up and have these as whole rooms, just like you have now with data servers, then there could be huge power savings," says Paul Roach at Loughborough University. There are other resources that biological data centres might need, such as nutrients to feed and keep alive the neuronal chips, but it should require far less cooling than conventional computing, says Roach. "The amount of energy that's saved with [Cortical Labs'] figures is fairly conservative." However, the technology is still at an early stage, says Tjeerd olde Scheper at Oxford Brookes University, who has worked with a competing biological computing company, FinalSpark. "Is it going to work as people might think? No, we're still in the early days of this development." It's hard to do a direct size comparison, as CL1 chips can't do conventional calculations like a regular silicon-based AI chip, but the proposed biological data centre will have hundreds of biological chips, compared to hundreds of thousands of graphics processing units (GPUs) seen in the largest AI data centres. "I think it is a very long way from production ready. It's a very big step from a small network playing a computer game to an LLM, " says Steve Furber at the University of Manchester. One of the remaining issues is that it's still unclear how to store the results of training these neurons in a form of memory or how to run actual computational algorithms on them, rather than training them for specific uses, like video games. Another challenge is how to retrain the neurons once they have completed a particular task. "Whatever they are trained on is lost when the culture ends its life, so there needs to be a proper retraining," says Scherer. "Then it's not an optimal solution to keep a technology going if you need to retrain every 30 days."

Human brain cells set to power two new data centers, thanks to 'body-in-the-box' CL1 -- Cortical Labs targets the AI energy crisis with biological computer that reportedly uses less energy than a calculator

Cortical Labs, the Australian tech startup behind the CL1 biological computer, has just announced a new partnership to establish data centers using its experimental product, Bloomberg reports. The biotech company has partnered with data center company DayOne to establish the two data centers, powered by the 'body-in-the-box' CL1s, in Australia and Singapore. This isn't the first we've seen from Cortical Labs in the last twelve months. The company has already appeared on our radar with the CL1's launch last March, with an update just a week ago that showcased its ability to play DOOM. Now, the Australian startup is looking at a more practical use for its product, thanks to these new data centers. The CL1 units consist of a silicon chip with 200,000 lab-grown human neurons 'grown' on top, converted initially from human blood cells. Like the complex array of cells in the human brain, these neurons respond to electrical stimuli to and from the chip, forming networks like the neural arrays in our own brains. The system is shown to be able to learn and adapt, with the CL1 equipped with life-support features that keep the conditions suitable for the neurons to remain alive for up to six months. A facility in Melbourne, Australia, will consist of 120 CL1 units, but it expects its joint project in Singapore with DayOne to have as many as 1,000 units, although "in phases," according to founder and CEO Hon Weng Chong. An initial 20 CL1 units will operate in an "initial validation phase" at the Yong Loo Lin School of Medicine at the National University of Singapore. From an industrial point of view, the CL1 could have potentially significant advantages over traditional computing, especially as companies search for ways to significantly increase capacity for AI without the huge increase in energy consumption. The neural approach used by the CL1 requires far less energy than a typical computer, according to Cortical Labs, requiring only a fraction of that used by conventional AI chips - less than a handheld calculator, even, according to Hon. These units are individually priced at around $35,000 each, based on Cortical Labs' launch details. Energy usage aside, Cortical Labs's experiments with Doom, and Pong before it, have demonstrated the ability of the CL1's neural networks to process data, learn from it, and adapt. This new collaboration with DayOne suggests work is ongoing to deploy the technology in a real-world environment that could serve as an alternative to power-hungry, water-demanding, and heat-generating AI data centers being built inside wind turbines and elsewhere. Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds.

Human Brain Cells Run New Data Centers in Singapore, Melbourne

Cortical Labs has made progress, including teaching its brain cells to play computer games like Pong and Doom, and is building facilities in Melbourne and Singapore to house its CL1 units, with the Singapore location set to deploy up to 1,000 units in phases. Biotech startup Cortical Labs is working on two small data centers run by human brain cells, putting lab-grown neurons onto silicon in an experiment that could one day challenge chips from the likes of Nvidia Corp. The Australia-based startup unveiled its first biological data center in Melbourne and is building another in Singapore with partner DayOne Data Centers Ltd., it said in a statement on Tuesday. Instead of racks of servers running on conventional processors, the facilities will house biological computers known as CL1 units, powered by human brain cells. While years or decades away from challenging mainstream technology, the project highlights scientists' search for novel solutions to address problems arising from an artificial intelligence-induced need for increasing amounts of computing capacity. The swift buildout of AI data centers across the planet has led to environmental concerns over their power needs and water consumption as well as shortages in silicon. The computing capacity of Cortical Labs' systems is modest, but the company is making progress. One of its earlier achievements was to teach its brain cells to play the rudimentary computer game Pong. Last month, it said it had trained them to play the much more advanced title Doom. The neurons used by Cortical Labs, grown from stem cells, sit on a chip that sends and receives electrical signals to the cells and records how they respond. This allows the company's software to interact with the cells and interpret their responses as computing output. Aside from the potential to tap into the brain's ability to hold and process data, there's an additional advantage to leveraging biology in this way: Neurons use very little energy. AI is driving a surge in electricity demand, forcing governments and technology companies to search for more efficient computing systems. Cortical Labs says its biological computers consume a fraction of the power used by conventional AI processors. Hon Weng Chong, the upstart's founder and chief executive officer, said in an interview that each CL1 unit uses less power than a handheld calculator. The Melbourne facility will house 120 CL1 units, while the Singapore location run with data center operator DayOne is set to deploy as many as 1,000 units in phases, Hon said. The Singapore project will start with an initial deployment at the National University of Singapore's Yong Loo Lin School of Medicine. The CL1 units use neurons that have been converted from human blood cells.

That Doom-running, human brain cell-powered computer is headed for data centers

Serving tech enthusiasts for over 25 years. TechSpot means tech analysis and advice you can trust. What just happened? Following news that its human brain cell-powered computer can run Doom, Australian biotech startup Cortical Labs has announced it is working on two small data centers running on the same technology. Cortical Labs made plenty of headlines last month when its latest hardware platform, the CL1, which uses living human neurons as the core of a fully functioning computer, was demonstrated running Doom. Now, the startup has announced a partnership with DayOne to build its first biological data center in Melbourne and another in Singapore. The facilities will use the CL1 biological computers instead of the usual processors found in conventional data centers. According to the latest report, the Melbourne site will consist of 120 CL1 units. The Singapore project is expected to be much larger, scaling to as many as 1,000 units in phases. An initial 20 CL1 units will be used in a validation phase at the Yong Loo Lin School of Medicine at the National University of Singapore. The hybrid biological computer pairs lab-grown human neurons with a silicon chip so software can communicate with living cells in real time. Cortical Labs says the self-contained system includes on-board life support, recording, and application handling, and can keep the neurons alive for up to six months. More recent reporting suggests the current CL1 setup uses around 200,000 neurons per unit, which is lower than some of the bigger figures attached to earlier DishBrain-era coverage. The company is also pitching the technology as a possible answer to the growing energy demands of AI infrastructure, with recent coverage claiming that each unit uses only a fraction of the power required by conventional AI chips. Cortical Labs CEO Hon Weng Chong even said the system uses less energy than a handheld calculator. As for how it works, the neurons grow across the chip and receive electrical stimuli that represent information from a simulated environment. Their responses are then read back into Cortical's Biological Intelligence Operating System, or biOS, creating a closed loop in which the cells can adapt to feedback in real time. The CL1 is less of a conventional processor than a programmable wetware platform: silicon handles the interfacing and control, while the biological component learns and self-organizes in ways standard hardware does not. That also means these data centers are not being positioned as direct replacements for traditional server farms, at least not yet. Right now, Cortical Labs appears to be making the case that biological computing could carve out a niche in areas where adaptability and lower energy use matter more than raw conventional performance.

The Company That Made a Dish of Neurons Play DOOM Is Getting Into Brain Cell-Powered Data Centers

A computer platform that runs on human neurons (and recently showed off said neurons' ability to play DOOM) now wants in on the data center boom. Australia-based Cortical Labs announced today that it is working on two of what it calls “biological data centers,†one in Melbourne and another in Singapore. Instead of relying on traditional GPU servers, these facilities would be powered by Cortical Labs’ biological CL1 units. Each of these bio-computers is equipped with a chip known as a “multi-electrode array†that has 200,000 human brain neurons grown directly on it. For comparison, the human brain is estimated to contain between 60 billion and 99 billion neurons. The announcement comes as tech giants like Google, Amazon, Microsoft, and OpenAI are spending billions of dollars to build new data centers to train and run their most advanced AI models. Beyond sounding like a mad scientist’s fever dream, a data center run on neurons could have a practical benefit. Traditional data centers are notoriously energy-hungry, with many requiring the equivalent power needed to supply entire neighborhoods. In some areas where large data centers have been built, the demand has strained local power grids and driven up electricity prices. That, in turn, has sparked community backlash against certain projects. Even President Donald Trump has gotten involved and made major tech companies pinky swear that they will cover the power costs of their projects. Conversely, neuron-powered computer chips surprisingly use far less energy than traditional AI chips. Cortical Labs CEO Hon Weng Chong told Bloomberg that each CL1 computer uses less power than a handheld calculator. The computers work by sending electrical signals to neurons grown from blood stem cells. The chip then records the neurons’ electrical responses to that stimulation as computing output. The company made headlines in February after posting a video showing how its team managed to get the neurons to play a version of the video game DOOM. Now the company is building a data center in Melbourne that will house 120 CL1 units. In Singapore, the company is partnering with sustainability-focused data center firm DayOne to set up a smaller prototype facility at the Yong Loo Lin School of Medicine at the National University of Singapore. The center will house a single rack of 20 CL1 units and serve as an “initial validation phase.†There are also plans to eventually deploy the technology in a commercial DayOne data center and test it under real-world conditions. The phased expansion could eventually scale up to deploy as many as 1,000 CL1 units at a DayOne facility. Cortical Labs and DayOne did not immediately respond to requests for comment from Gizmodo.