Innovations in Microchip and Semiconductor at HUST

In late February 2025, Assoc. Prof. Nguyen Duc Minh, Vice Dean of the School of Electrical and Electronic Engineering at Hanoi University of Science and Technology (HUST), and the EDABK research team won first place in the international chip design competition organized by Efabless and co-sponsored by Google. Their design marks the first neuromorphic chip, an AI-based architecture simulating human brain activity, by HUST.

Compete for the chance to build a chip for free

Assoc. Prof. Nguyen Duc Minh and the EDABK research team from the BKIC Lab (a lab for IC and embedded systems design) at HUST developed a neuromorphic chip based on Spiking Neural Networks (SNN), an AI architecture introduced in 1997 known for its ultra-low power consumption, making it ideal for battery-powered and edge computing devices. SNNs also enable fast processing and better simulation of biological neural processes, supporting real-time applications.

The current chip prototype simulates 32 hardwired neurons and 1,000 synapses, with software extensions allowing up to 1,000 soft neurons and 65,000 soft synapses. Consuming only 1 mW of power, the chip is suitable for wearables like heart rate monitors for arrhythmia and stroke detection. It also achieves 96% accuracy in handwriting recognition.

The team leveraged AI to streamline the chip design process, reducing development time from concept to prototype to just 12 days, built on foundational research that began in 2019.
 

The design earned the EDABK team first prize in the international chip design competition organized by Efabless and co-sponsored by Google. As champions, the team won the opportunity to have their AI chip fabricated for free.

Reflecting on the biggest challenge, Prof. Minh shared: “The hardest part was we had no funding.” He explained that financial constraints prevented the team from accessing commercial PDK libraries and modern design tools, forcing them to rely entirely on open-source platforms. Winning the competition was therefore a critical breakthrough, allowing the team to prototype their chip without cost.

Looking ahead, the EDABK team plans to expand the chip’s capabilities by simulating more neurons and mimicking more complex brain functions with higher accuracy. They are also exploring practical applications to bring the product into real-world use.

Aiming to control the full production process of next-gen LEDs

Besides the School of Electrical and Electronic Engineering, other units at HUST are also advancing semiconductor research across different industry stages. Notably, Assoc. Prof. Dao Xuan Viet’s team from the School of Materials Science and Engineering leads cutting-edge work on specialized LED technology for agriculture and human-centric lighting.

Amid rapid global and local semiconductor development, Assoc. Prof. Viet’s group focuses on phosphor materials and advanced LED packaging technology, collaborating closely with businesses for technology transfer.

Their work targets two main applications: high-tech agriculture and human-centric lighting. For agriculture, they successfully developed red, far-red, and unique dual-emission purple-red LEDs that stimulate off-season flowering, boosting economic value and crop control for farmers.

In human-centric lighting, the team optimized emission spectra to create LEDs that biologically suppress melatonin, enhancing alertness and productivity. They also developed cyan and low-blue light LEDs, promoting healthier lighting ecosystems.
 

Building on strong expertise in LED packaging, the group is expanding into semiconductor materials and LED chip development, ensuring seamless integration from research to production. This marks a crucial step toward mastering the full manufacturing process of next-generation LEDs.