Artificial intelligence has advanced rapidly over the past decade, but researchers continue to search for the elusive breakthrough known as artificial general intelligence (AGI). Unlike today’s AI tools, which specialize in narrow tasks, AGI would mimic the flexibility of human intelligence—adapting to a wide range of challenges. In a major step toward that vision, Chinese scientists have unveiled the world’s largest neuromorphic supercomputer, nicknamed “Darwin Monkey” or “Wukong.”
Developed at Zhejiang University in collaboration with Zhejiang Lab and backed by Alibaba Group, this cutting-edge machine is modeled on the neural structure of the macaque monkey brain. With more than 2 billion artificial neurons and over 100 billion synapses, Darwin Monkey represents one of the boldest attempts yet to bridge the gap between biological intelligence and machine computation.
What Makes Darwin Monkey Different?
Unlike conventional supercomputers that rely on raw processing power and traditional silicon-based architectures, Darwin Monkey is built on spiking neural networks (SNNs). These networks closely replicate how neurons communicate in the human and animal brain. Instead of continuously processing binary states of 0s and 1s, SNNs transmit information through spikes of electrical activity—firing only when a sufficient input threshold is reached.
This approach allows the system to:
- Process data in parallel, similar to the human brain.
- Use energy more efficiently, avoiding constant high-power demands.
- Simulate brain-like functions such as perception, problem-solving, and logical reasoning.
Remarkably, the entire supercomputer draws just 2,000 watts of power—roughly the energy needed for a couple of household appliances. This efficiency comes from 960 Darwin III neuromorphic chips, each containing millions of spiking neurons.
From Darwin Mouse to Darwin Monkey: A Growing Lineage
Darwin Monkey did not appear overnight. It follows a steady series of advances in China’s neuromorphic research.
- 2020 – Darwin Mouse: This earlier system modeled the brain of a rodent, simulating around 120 million artificial neurons. It proved that neuromorphic chips could replicate complex biological behaviors on a smaller scale.
- 2023 onward – Expanded focus: Researchers began applying these systems to simulate simpler animal brains, including zebrafish and mice, laying the groundwork for scaling up to more advanced structures.
- 2025 – Darwin Monkey: With its billions of neurons and unprecedented scale, the system is now capable of supporting logical reasoning, mathematical problem-solving, and even generating content.
Each generation builds on lessons learned from the last, showing a roadmap toward machines that more closely mirror human cognition.
Why Neuromorphic Computing Matters
The rise of neuromorphic computing reflects a broader realization: simply adding more processing cores to conventional supercomputers will not automatically create intelligence.
Traditional supercomputers excel at tasks like weather forecasting, simulating nuclear reactions, or running large-scale physics models. But intelligence is not just raw power—it’s about adaptability, efficiency, and parallel processing. The human brain, with its roughly 86 billion neurons, operates on just 20 watts of power—less than a dim light bulb.
Neuromorphic machines like Darwin Monkey aim to mimic this model by emulating structure as much as function. Instead of brute-force calculation, they process information the way brains naturally do, making them ideal for applications in:
- Artificial general intelligence research
- Robotics and autonomous systems
- Cognitive science and neuroscience studies
- Healthcare innovations, such as simulating brain disorders
- Content generation and creative AI tools
Comparisons to Global Competitors
China’s breakthrough with Darwin Monkey arrives in a field that remains relatively young. Other global players have made significant progress:
- Intel’s Hala Point previously held the record, boasting 1.15 billion artificial neurons. While impressive, it falls short of Darwin Monkey’s 2 billion neurons.
- IBM’s TrueNorth chip, developed in the mid-2010s, pioneered early neuromorphic designs but remains limited in scale compared to today’s systems.
- European Union projects under the Human Brain Project have sought to digitally simulate brain structures, though they often rely on conventional architectures rather than spiking neural models.
Comparisons can be difficult, as architectures differ in design and goals. However, Darwin Monkey’s scale, efficiency, and focus on biological realism place it among the most ambitious efforts worldwide.
Potential Applications and Early Use Cases
Although still experimental, Darwin Monkey is already being used for groundbreaking research. Scientists have applied it to:
- Model animal brains (zebrafish, mice, and macaques) for deeper insights into cognitive science.
- Study logical reasoning in machines, providing a foundation for AGI.
- Explore creative content generation, potentially revolutionizing industries such as media, design, and entertainment.
- Advance mathematical problem-solving by running parallel simulations at efficiency levels traditional systems can’t match.
In the long run, neuromorphic systems like Darwin Monkey may support applications in real-time language learning, vision-based navigation for autonomous vehicles, and even advanced medical diagnostics.
Challenges and Limitations
Despite its promise, Darwin Monkey is far from delivering true human-like intelligence. Key challenges remain:
- Complexity of the brain: Even with billions of neurons, the human brain’s structure and interactions are vastly more intricate.
- Software ecosystem: Developing algorithms optimized for spiking neural networks remains a significant hurdle.
- Scalability: While Darwin Monkey sets a new record, scaling further toward human brain-level complexity will demand new breakthroughs in chip design and energy management.
- Ethical considerations: As machines become more brain-like, questions arise about responsibility, decision-making, and the societal impact of AGI.
Researchers acknowledge these hurdles but remain optimistic that neuromorphic computing provides a viable path forward.
China’s Strategic Position in AI and Supercomputing
China’s unveiling of Darwin Monkey also carries geopolitical significance. The country has invested heavily in AI, quantum computing, and next-generation semiconductors, seeking to establish technological independence and global leadership.
By surpassing existing international benchmarks, Darwin Monkey demonstrates China’s ability to innovate beyond replication, carving its place as a key player in the race for AGI. With backing from tech giants like Alibaba and strong university research partnerships, China is positioning neuromorphic computing as a national priority.
Frequently Asked Questions:
What is China’s Darwin Monkey supercomputer?
Darwin Monkey, also known as Wukong, is the world’s largest neuromorphic supercomputer, designed to mimic the structure and function of the brain using spiking neural networks.
How powerful is Darwin Monkey compared to other neuromorphic systems?
It contains more than 2 billion artificial neurons and over 100 billion synapses, surpassing Intel’s Hala Point, which held the previous record with 1.15 billion neurons.
How does Darwin Monkey differ from traditional supercomputers?
Unlike conventional systems that rely on raw computational cores, Darwin Monkey processes information using brain-inspired neural spikes, making it more energy-efficient and biologically realistic.
What is the energy consumption of Darwin Monkey?
The entire system consumes around 2,000 watts of power, significantly lower than most supercomputers of its scale.
Who developed Darwin Monkey?
It was developed by Zhejiang University and Zhejiang Lab in collaboration with Alibaba Group.
What are the early applications of Darwin Monkey?
Researchers are using it for simulating animal brains, studying cognitive science, exploring logical reasoning, mathematical problem-solving, and even creative content generation.
Why is neuromorphic computing important?
Neuromorphic systems are designed to function like biological brains, offering higher efficiency, adaptability, and parallel processing—crucial for the development of artificial general intelligence (AGI).
Conclusion
China’s unveiling of the Darwin Monkey supercomputer marks a turning point in neuromorphic computing and the global pursuit of artificial general intelligence (AGI). With its record-breaking 2 billion neurons, 100 billion synapses, and remarkable energy efficiency, the system demonstrates how brain-inspired designs can reshape the future of AI. While challenges remain in scaling and software development, Darwin Monkey provides a groundbreaking platform for cognitive science, AI research, and real-world applications. As global competition intensifies, this achievement underscores China’s growing leadership in next-generation computing and the quest to build machines that think more like humans.
