Scientists from Peking University and Beijing University of Posts and Telecommunications have made a major breakthrough in chip technology. They have created the world’s first AI chip using carbon nanotubes (CNTs), a material with excellent electrical and mechanical properties.
Unlike traditional silicon chips, this new chip uses ternary logic instead of the usual binary system. This means it processes data in three states instead of just zeros and ones. This makes computing faster and more energy-efficient.
The team developed a special carbon nanotube transistor using a method called source-gated transistors (SGTs). This allows the chip to switch between three different electrical states, improving speed and reducing power use.
To test their new chip, the researchers built an AI neural network that mimics the human brain. The network successfully recognized handwritten numbers with perfect accuracy. This shows the chip’s potential for AI applications, image recognition, and machine learning.
Lead researcher Peng Lianmao, a member of the Chinese Academy of Sciences, has been working on carbon-based chips for over 20 years. In 2020, his team created an eight-inch CNT wafer, which performed better than similar silicon-based chips.
The new chip is highly stable, resistant to interference, and energy-efficient. It can be used in AI, high-performance computing, and IoT (Internet of Things) devices. However, CNT chips are still behind silicon chips in terms of transistor density.
For example, Nvidia’s RTX 5090 GPU, announced in 2025, has 92 billion transistors, far more than current CNT chips. Still, carbon-based chips are seen as the future of semiconductor technology, and China is leading this development.
Peng Lianmao believes that carbon chips could become mainstream in 10 to 15 years, replacing silicon chips in smartphones, supercomputers, and data centers. This could start a new era of faster, more efficient computing.
Source: tribune
