WUJI: The World’s First 2D Semiconductor Chip
20 April 2025 · Uncategorized ·
As global challenges arise from the approaching physical limits of Moore’s Law, two-dimensional (2D) semiconductors—materials with a single atomic layer thickness—are internationally recognized as key to overcoming these limitations. Scientists are actively exploring their application in integrated circuits (ICs). For over a decade, both academic and industrial sectors worldwide have mastered wafer-scale growth techniques for 2D materials, successfully manufacturing high-performance basic devices measuring hundreds of atoms long and just several atomic layers thick. However, prior to the breakthrough by Fudan University’s team led by Zhou Peng from the School of Microelectronics in April 2025, the highest integration level achieved internationally was only 115 transistors for a two-dimensional semiconductor digital circuit—a milestone accomplished by Vienna Technical University's team in 2017. A core challenge has been assembling these atomically precise components into complete IC systems while maintaining process precision and uniformity.
On April 2nd, 2025, Zhou Peng (recipient of the ‘Science Exploration Award’ for Information Electronics) and his joint research group from Fudan University published a paper titled “A RISC-V 32-Bit Microprocessor Based on Two-dimensional Semiconductors” in *Nature*. This team overcame integration density bottlenecks to successfully develop 'WUJI' ('Infinite'), the world’s first microprocessor based on two-dimensional semiconductor materials (molybdenum disulfide, MoS₂). Under control of 32-bit input instructions, “WUJI” can perform up to four billion data addition/subtraction operations and supports GB-level data storage/access. It also allows for program writing with a maximum length of one hundred million simplified instruction sets.
The name 'WUJI' (Infinite) symbolizes creation from nothingness without limits—reflecting its status as currently the largest electronic circuit built using 2D semiconductors, containing over five thousand nine-hundred molybdenum disulfide field-effect transistors and featuring a logic path composed of seventeen cascaded logical elements between flip-flops. The system operates at four volts with external clock signal control, functioning independently without additional bias or control signals.
The fabrication process involves front-end-of-line (FEOL) processes for the source/drain layers and gate layers containing working transistors, followed by back-end-of-line (BEOL) processes to form logic connection layers and module interconnection layers. The team addressed core challenges in scaling up 2D materials while avoiding leakage currents and stability issues through ‘atomic interface precision control’ technology combined with AI-driven process optimization.
In terms of chip architecture design, the Fudan University research group selected an open-source RISC-V instruction set along with AI-assisted design tools to explore a new path free from traditional silicon-based patent constraints. This approach not only lowers technical barriers but also paves the way for building autonomous and controllable chip ecosystems.
The performance potential of 2D semiconductor chips lies primarily in two areas: power consumption efficiency breakthroughs due to quantum confinement effects, enabling efficient operation at extremely low voltages—making them suitable for edge computing scenarios where battery life is critical; and their atomic-level thickness allows vertical stacking beyond the planar integration limits of traditional silicon-based chips.
Furthermore, 2D materials offer unique advantages in quantum computing interfaces: they can serve as a bridge between qubits and classical circuits through interface control to achieve precise manipulation of quantum states for hybrid quantum-classical computation architectures. These characteristics make them key drivers pushing AI computational power from centralized cloud servers towards distributed terminal devices, reducing latency while enhancing privacy protection.
Despite these advancements, 2D semiconductors are not expected to replace silicon-based chips but rather complement their technology paths, as Zhou Peng noted: “Just like the emergence of subways did not render buses obsolete; two-dimensional semiconductor and silicon chip technologies coexist in a complementary relationship.”
Currently, efforts focus on enhancing performance and integration density while bridging with existing silicon production lines for industrialization. The team believes that within a short period, 2D semiconductors can achieve comparable performance to traditional chips.
However, transitioning from lab prototypes like 'WUJI' towards full-scale commercialization still faces significant hurdles including insufficient technological maturity, lack of supply chain support, and resistance due to existing market ecosystems; the team is working on improving integration density while collaborating with industry partners for practical applications.
On April 2nd, 2025, Zhou Peng (recipient of the ‘Science Exploration Award’ for Information Electronics) and his joint research group from Fudan University published a paper titled “A RISC-V 32-Bit Microprocessor Based on Two-dimensional Semiconductors” in *Nature*. This team overcame integration density bottlenecks to successfully develop 'WUJI' ('Infinite'), the world’s first microprocessor based on two-dimensional semiconductor materials (molybdenum disulfide, MoS₂). Under control of 32-bit input instructions, “WUJI” can perform up to four billion data addition/subtraction operations and supports GB-level data storage/access. It also allows for program writing with a maximum length of one hundred million simplified instruction sets.
The name 'WUJI' (Infinite) symbolizes creation from nothingness without limits—reflecting its status as currently the largest electronic circuit built using 2D semiconductors, containing over five thousand nine-hundred molybdenum disulfide field-effect transistors and featuring a logic path composed of seventeen cascaded logical elements between flip-flops. The system operates at four volts with external clock signal control, functioning independently without additional bias or control signals.
The fabrication process involves front-end-of-line (FEOL) processes for the source/drain layers and gate layers containing working transistors, followed by back-end-of-line (BEOL) processes to form logic connection layers and module interconnection layers. The team addressed core challenges in scaling up 2D materials while avoiding leakage currents and stability issues through ‘atomic interface precision control’ technology combined with AI-driven process optimization.
In terms of chip architecture design, the Fudan University research group selected an open-source RISC-V instruction set along with AI-assisted design tools to explore a new path free from traditional silicon-based patent constraints. This approach not only lowers technical barriers but also paves the way for building autonomous and controllable chip ecosystems.
The performance potential of 2D semiconductor chips lies primarily in two areas: power consumption efficiency breakthroughs due to quantum confinement effects, enabling efficient operation at extremely low voltages—making them suitable for edge computing scenarios where battery life is critical; and their atomic-level thickness allows vertical stacking beyond the planar integration limits of traditional silicon-based chips.
Furthermore, 2D materials offer unique advantages in quantum computing interfaces: they can serve as a bridge between qubits and classical circuits through interface control to achieve precise manipulation of quantum states for hybrid quantum-classical computation architectures. These characteristics make them key drivers pushing AI computational power from centralized cloud servers towards distributed terminal devices, reducing latency while enhancing privacy protection.
Despite these advancements, 2D semiconductors are not expected to replace silicon-based chips but rather complement their technology paths, as Zhou Peng noted: “Just like the emergence of subways did not render buses obsolete; two-dimensional semiconductor and silicon chip technologies coexist in a complementary relationship.”
Currently, efforts focus on enhancing performance and integration density while bridging with existing silicon production lines for industrialization. The team believes that within a short period, 2D semiconductors can achieve comparable performance to traditional chips.
However, transitioning from lab prototypes like 'WUJI' towards full-scale commercialization still faces significant hurdles including insufficient technological maturity, lack of supply chain support, and resistance due to existing market ecosystems; the team is working on improving integration density while collaborating with industry partners for practical applications.