News

April 15, 2024

Chiplet Technology could transform everyday AI – Peter Esenju Mbua

Chiplet Technology could transform everyday AI – Peter Esenju Mbua

By Kenneth Oboh

Peter Esenju Mbua, a graduate research assistant at the University of Florida, has said his work in Chiplet Technology could transform how everyday devices use artificial intelligence, AI, especially at the edge where speed, security, and efficiency matter most.

Mbua, who works with the Electrical and Computer Engineering Department, is part of the National Science Foundation’s Central Florida Semiconductor Innovation Engine (Award No. 1946088). He is investigating how processors built with chiplets can support AI in tools ranging from smartwatches and home gadgets to medical devices and self-driving cars.

“AI is no longer confined to research labs or major tech firms. It is embedded in our daily lives, whether through voice assistants, wearables, or autonomous vehicles. But current edge devices face serious limits, from battery drain to security risks. That is where chiplet technology comes in,” Mbua said.

The project, led by principal investigator Dr. Christophe Bobda with Mbua as a key researcher, is rethinking how processors are designed. Instead of relying on one large, general-purpose chip, chiplet technology allows engineers to assemble smaller, specialized parts like building blocks. This modular approach makes it possible to tailor processors for specific tasks, whether monitoring heart health or analyzing road conditions in real time.

According to the team, the benefits for consumers could be far-reaching. Devices could enjoy longer battery life, use less energy, and provide stronger protection from hackers. Everyday tools such as fitness trackers, smart home systems, and security devices would also run faster and more reliably.

The research is already at an advanced stage. Using high-level computer modeling tools, the team is building simulators to test chiplet-based designs before real-world deployment. This approach helps reduce costs and provides early proof of concept.

In the long term, Mbua believes this work will not only make AI-powered devices smarter and more sustainable but also strengthen U.S. leadership in semiconductor innovation and cybersecurity.

“This is about making the devices we use every day not just smarter but also more secure, energy-efficient, and ready for the future,” he said.