Researchers from Zhejiang University in China have announced the development of Darwin Monkey, which is currently the largest brain-inspired computer built to date. Developed by the State Key Laboratory of Brain-Machine Intelligence, this system represents a significant step forward in neuromorphic computing, aiming to create machines that mimic how biological brains work.
Neuromorphic computing is a field that designs computer systems inspired by the structure and function of the human brain. Unlike traditional computers that process data in a linear, step-by-step manner, neuromorphic systems use hardware that imitates neural networks—networks of interconnected nerve cells (neurons) in the brain. This approach allows for more efficient, faster, and adaptable processing, similar to how our brains operate.
Darwin Monkey, also called Wukong, is the first neuromorphic computer built entirely with specialized chips designed to simulate neural activity. It can model over 2 billion neurons, roughly equivalent to the number of neurons in a macaque monkey’s brain—a common animal used in neuroscience because its brain structure resembles that of humans. This achievement also surpasses previous records including Intel’s Hala Point system, released in 2024, which was capable of simulating over 1 billion neurons.
The supercomputer consists of 960 chips called Darwin 3—the third generation of brain-inspired processors. These chips are assembled into 15 blade-shaped servers. Each chip supports more than 2.35 million neurons that send electrical signals (called spikes) and connects through hundreds of millions of synapses—the connections that enable neurons to communicate.
Despite its enormous scale, the entire system consumes about 2,000 watts of power (roughly the amount used by a small home appliance), demonstrating impressive energy efficiency for such a powerful machine.
Inspiration from the Brain
The driving idea behind Darwin Monkey is to replicate the brain’s incredible efficiency and ability to process many tasks at once. Unlike traditional computers that rely on linear processing and often consume a lot of energy, neuromorphic systems mimic the brain’s parallel neural networks, leading to low power consumption, high speed, and adaptive learning.
The team developed a special set of instructions tailored for brain-inspired computing, along with a learning mechanism that allows the system to adapt and improve in real time. They also created a new operating system specifically designed for this type of hardware, enabling scientists to run advanced AI models directly on the supercomputer.
Using this platform, researchers have successfully tested large neural models, including a brain-like model called DeepSeek, which can perform reasoning and generate content. The system can also simulate smaller animal brains—such as those of C. elegans (a tiny worm), zebrafish, mice, and macaques—providing new tools for neuroscience research.
This development opens exciting possibilities across many fields. Its large scale and ability to process many tasks simultaneously could lead to more efficient artificial intelligence (AI) systems that better mimic human reasoning. It also provides neuroscientists with a powerful tool to study brain functions without invasive experiments.
As research advances, these systems might redefine the future of computing, neuroscience, and artificial intelligence.
Leave a Reply