Корпус CL1 містить системи життєзабезпечення, необхідні для виживання клітин людського мозку / Cortical Labs
Cortical Labs‘ CL1 Synthetic Biological Intelligence (SBI), the world’s first computer to combine human brain cells with conventional chips to form a neural network, has been launched into commercial production.
The CL1 system was officially unveiled in Barcelona on March 2. It is expected to revolutionize scientific and medical research. Neural networks based on human cells formed on a silicon chip make up an organic computer that is constantly learning. The company’s engineers say that it learns so quickly and flexibly that it far outpaces silicon chips for artificial intelligence calculations used to train large language models (LLM) programs, such as ChatGPT.
«We offer [service]«Wetware-as-a-Service» (WaaS). Customers will be able to buy the CL1 biocomputer outright or simply purchase time on the chips with remote access to work with cultured cell technology via the cloud. This platform will enable millions of researchers, innovators, and thinkers around the world to turn the potential of the CL1 into tangible, real-world impact. We will provide them with the platform and support to invest in research and development and drive new breakthroughs and research,» says Cortical Labs founder and CEO Dr. Hong Weng Chong.
The Cortical Labs team made headlines in 2022 after developing a self-adaptive computer «brain» with 800 thousand human and mouse neurons on a chip, and taught this network to play a video game. CL1 computers will be widely available in the second half of 2025. New Atlas journalists visited the company’s headquarters in Melbourne and say that «the potential here is much greater than [playing] Pong».
«We almost see it as a kind of other life form, say an animal or a human. We think of it as a mechanical and engineering approach to intelligence. We’re using the substrate of intelligence, which is biological neurons, but we’re assembling them in a new way,» said the company’s chief scientist Brett Kagan in 2023.
Cortical Labs has come a long way since DishBrain — an important but outdated first step. Now that CL1 has been commercialized, scientists can get hands-on with the technology and begin exploring a wide range of real-world applications.
«The current version — it’s a completely different technology. The previous one used a CMOS chip, which essentially gave you a very dense reading, but it was opaque, you couldn’t see the cells. There were other problems, too: for example, when you stimulate [the process] with a CMOS chip, you can’t take out the charge; you can’t balance the charge either. You end up with a charge buildup where you stimulate for a long time, and that’s very bad for the cells. … This version has much simpler technology, but that means it’s much more stable, and you have much more ability to actively balance,» Kagan told reporters during their visit in late 2024.
Inside the CL1, laboratory-grown neurons are placed on a flat electrode array — «generally just metal and glass». The 59 electrodes form the basis of a more stable network and offer the user a high degree of control over the activation of the neural network. This «brain» is placed in a rectangular life support unit that is connected to a system based on special software that operates in real time.
«A simple way to describe it would be a body in a box, but it has wave filtering, it has a place where the data is stored, it has pumps that provide circulation, mixing of gases, and of course temperature control,» Kagan explained.
Cortical Labsl is assembling these computers into blocks to build a first-of-its-kind biological neural network server stack containing 30 separate blocks, each containing cells on its own array of electrodes. the online launch is expected in the coming months. The team plans to launch four such stacks, which will be available for commercial use via a cloud-based system by the end of the year. The individual devices are expected to cost about $35,000 thousand.
The CL1 unit stack consumes only about 850-1000 watts of power, is fully programmable, and offers «bidirectional stimulation and a reading interface adapted to enable neural communication and network training», the team noted in its presentation. The CL1 unit does not require an external computer to operate. Related article New Atlas contains a lot of additional data about the system, statements by scientists, in particular, on the ethical and investment sides of the technology.