BEIJING, 2026-02-01 06:41:00
Flexible AI Chips Could Revolutionize Wearable Tech
Table of Contents
Researchers in China have developed a new family of flexible chips poised to overcome limitations in current smart wearable technology.
- Scientists at Tsinghua University and Peking University created FLEXI, a flexible AI chip.
- The chip processes data directly, eliminating the need for external computers.
- FLEXI demonstrated 99.2% accuracy in identifying irregular heartbeats during testing.
- The chip consumes less than 1% of the energy of standard chips and costs less than $1 to mass produce.
Smart wearables promise a lot, but often fall short of their potential. A key roadblock? The rigid, power-hungry, and brittle chips that power them. Now, researchers are aiming to change that. A team from Tsinghua University and Peking University in China has developed FLEXI, a family of flexible chips designed to overcome these limitations.
A Flexible Solution
Details of the chip’s design, along with its ability to handle complex AI tasks—including processing data from body sensors to identify health issues like irregular heartbeats in real time—were published in the journal Nature.
Unlike current wearables that typically send data to an external computer for processing, this chip can process information directly on the device. AI is integrated into the chip’s circuitry, meaning memory components perform calculations on stored data. This eliminates data transfer delays and conserves power.
Real-World Results
Scientists tested FLEXI in real-world health scenarios, collecting data from a group of volunteers. The chip accurately identified irregular heartbeats in 99.2% of cases and tracked daily movements, including cycling and walking, with 97.4% accuracy.
During research, the chip consumed less than 1% of the energy of a standard chip, and is inexpensive to produce. FLEXI costs less than a dollar during mass production. It’s a robust, cost-effective solution that withstands mechanical stress and ensures long-term operational stability.
