Chinese Scientists Develop World’s First Carbon-Based AI Chip
- Fahad Bin Khalid
- March 10, 2025
- 5:37 pm
- 160
- Technology
A New Era in Computing
Chinese scientists have developed the world’s first carbon-based AI chip, marking a major breakthrough in semiconductor technology. This innovation could revolutionize artificial intelligence and computing by making chips faster and more energy-efficient.
How the Carbon-Based AI Chip Works
Unlike traditional silicon chips, this carbon-based AI chip uses carbon nanotubes (CNTs). These nanotubes are thin, strong, and excellent conductors of electricity. Scientists have used them to build a new type of transistor that allows for better performance than silicon-based chips.
Ternary Logic: A Game Changer
Most computers use binary logic, meaning they process information using only ones and zeros. However, this carbon-based AI chip uses ternary logic, which adds a third state to data processing. This approach makes computing faster and more efficient while using less energy.
Neural Network Testing and AI Applications
Researchers tested the chip by creating a neural network that mimics human brain activity. The chip successfully classified handwritten digits with perfect accuracy, proving its potential for machine learning and AI applications. This technology could improve image recognition, speech processing, and advanced computing tasks.
China’s Leadership in Carbon Nanotube Technology
Professor Peng Lianmao, a leading expert in carbon-based semiconductors, has worked on this technology for over 20 years. His team created an eight-inch CNT wafer in 2020, showing that carbon nanotube transistors can outperform silicon-based ones in certain areas.
Future Applications and Challenges
This chip offers high stability and resistance to interference, making it ideal for AI, machine learning, and Internet of Things (IoT) devices. However, carbon nanotube chips still have lower integration density compared to traditional chips. For example, Nvidia’s RTX 5090 GPU has 92 billion transistors, far more than what CNT technology can currently achieve.
The Road Ahead
Scientists believe that carbon-based AI chips will become mainstream within 10 to 15 years. If successful, they could replace silicon chips in supercomputers, smartphones, and data centers. This shift would lead to faster, more energy-efficient devices, shaping the future of computing technology.
