Chinese engineers have unveiled the world’s first brain‑like computer with over 2 billion artificial neurons. Named Darwin Monkey (also called Wukong), the system was developed at Zhejiang University’s State Key Laboratory of Brain‑Machine Intelligence and represents a major leap in neuromorphic computing
Background: What Is a Brain‑Like Computer?
Brain‑like computer, also known as a neuromorphic system, mimics the structure and function of the biological brain by using artificial neurons and synapses. Unlike traditional linear processors, these systems operate in parallel, use pulsed signals (spikes), and can adaptively learn in real time with far less energy consumption
Development: Meet Darwin Monkey (Wukong)
Zhejiang University’s team built Darwin Monkey using 960 Darwin 3 chips, each supporting about 2.35 million spiking neurons, for a total exceeding 2 billion neurons and more than 100 billion synapses. The entire system consumes just around 2,000 watts—about the power of a small home appliance.
This is the first system to combine large-scale neural simulation with integrated perception and cognition capabilities—including vision, hearing, language, and learning functions—achieved using the DeepSeek brain‑like large model
Significance: What Makes This a Milestone?
- Unmatched scale: Surpasses Intel’s Hala Point system, which previously simulated 1.15 billion artificial neurons in 2024
- Energy efficiency: Uses only 2,000 W versus Intel’s 2,600 W footprint, yet with higher neuron capacity.
- Advanced functionality: Performs tasks like reasoning, content generation, and mathematical problem-solving using AI models.
Applications and Future Directions
AI and Neuroscience
Darwin Monkey enables real-time AI reasoning and content generation via DeepSeek, plus simulations of animal brains ranging from worms to macaques. This opens new paths for neuroscience research without invasive procedures
Broader Use Cases
Potential applications span across intelligent edge devices, energy-efficient AI agents, real-time learning systems, and large-scale brain-modeling for research and commercial use.
How This Compares to Intel’s Hala Point
| System | Neurons | Synapses | Power Usage | Year |
|---|---|---|---|---|
| Darwin Monkey | >2 billion | ~100 billion | ~2 kW | 2025 |
| Hala Point | ~1.15 billion | ~128 billion | ~2.6 kW | 2024 |
Darwin Monkey not only doubles neuron capacity but also incorporates richer AI functions within a leaner energy profile.
Future Outlook
Researchers aim to further scale such neuromorphic systems, adding capabilities for more complex cognition, adaptability, and broader integration with AI tools. This technology may lead to more brain‑like large language models, smarter robotics, and efficient AI ecosystems.
