China’s Darwin Monkey: World’s First Brain-Like Supercomputer Rivaling Monkey Brain Complexity

Chinese engineers at Zhejiang University have unveiled the Darwin Monkey, the world’s first brain-inspired supercomputer built on neuromorphic architecture featuring over 2 billion artificial neurons and more than 100 billion synapses. This leap brings computing power close to the scale of a macaque monkey brain, a breakthrough that may reshape the future of artificial general intelligence (AGI), brain simulation, and low-power AI systems.

The Darwin Monkey system is powered by 960 Darwin 3 neuromorphic chips, a result of collaborative development between Zhejiang University and Zhejiang Lab, a research institute backed by the Zhejiang provincial government and Alibaba Group. Each Darwin 3 chip supports up to 2.35 million pulsed or spiking neurons, using specialized instruction sets designed to emulate brain-like learning and reasoning. In aggregate, the Darwin Monkey mimics the structure of a small primate brain while consuming just 2,000 watts of power, showcasing a new frontier in low-power, massively parallel computing.

Unlike traditional artificial neural networks that use continuous values, neuromorphic systems like Darwin Monkey operate through spiking neural networks (SNNs), firing discrete electrical-like spikes to process and transmit data. This model is far closer to biological neurons and opens up possibilities for more energy-efficient and adaptive AI.

According to Zhejiang University’s announcement on Chinese social media, the Darwin Monkey has successfully completed complex cognitive tasks such as content generation, logical reasoning, and mathematical computations, using a large-scale AI model developed by Chinese AI startup DeepSeek. Beyond industrial AI, its neuron-scale computing capabilities are now being leveraged to simulate entire brains of various animals, including zebrafish, mice, and monkeys, offering valuable tools for neuroscience and biomedical research.

This unveiling marks the next evolution in the university’s roadmap, following the 2020 release of the Darwin Mouse, which simulated 120 million neurons. By contrast, the new Darwin Monkey increases neuron count by over 16 times, and incorporates breakthroughs in chip interconnectivity, neuromorphic OS development, and online learning mechanisms, according to Pan Gang, director of the State Key Laboratory of Brain-Computer Intelligence.

The emergence of Darwin Monkey positions China as a serious contender in neuromorphic computing, a field long dominated by U.S. firms. Notably, Intel’s Hala Point, announced in April 2023, was previously the most advanced neuromorphic system, housing 1.15 billion neurons and installed at Sandia National Laboratories. China’s Darwin Monkey now surpasses that figure.

Experts believe brain-inspired architectures could redefine how the world approaches AI scaling, as traditional GPU-based systems like those from NVIDIA hit energy, memory, and latency walls. For investors and policymakers, neuromorphic chips may be the next battleground in sovereign AI infrastructure, especially given their importance to military, robotics, and edge computing applications.