» Current | 2021 | 2020 | 2019 | Subscribe

Linley Newsletter

Expedera Tops DLA-IP Charts

April 20, 2021

Author: Linley Gwennap

Targeting a rapidly growing edge-AI market, startup Expedera has developed a unique deep-learning-accelerator (DLA) architecture that provides far greater performance per watt than existing designs. Rather than using a compiler, the Origin architecture natively executes many convolutional neural networks (CNNs), eliminating the complexities of fetching and executing instructions. The startup, which emerged from stealth today at the Linley Spring Processor Conference, licenses its intellectual property (IP) to SoC developers. Using a 7nm test chip, it has validated the architecture’s power and performance at up to 18 trillion operations per second (TOPS) per watt.

AI-enabled edge devices—including smartphones, wearables, smart-home products, and automobiles—totaled about 600 million units in 2020, according to a recent report from The Linley Group that forecasts growth to 2.0 billion units in 2025. The Expedera IP can scale from 1 to 100 TOPS, making it well suited to applications from simple doorbell cameras to high-resolution cameras in self-driving cars.

Former Cisco employees Da Chuang, Sharad Chole, and Siyad Ma founded the company in 2018. CEO Chuang and VP Engineering Ma have extensive ASIC-design experience, while Chief Scientist Chole is a software architect with AI expertise. Chuang formerly cofounded Memoir Systems, which licensed high-bandwidth SRAM IP suitable for networking chips; Cisco acquired the startup in 2015. Expedera has raised undisclosed funding from strategic and angel investors. The small startup has signed a licensing deal with a top smartphone-chip company, which is already using the Origin IP in its SoC development.

Subscribers can view the full article in the Microprocessor Report.

Subscribe to the Microprocessor Report and always get the full story!

Purchase the full article

Free Newsletter

Linley Newsletter
Analysis of new developments in microprocessors and other semiconductor products
Subscribe to our Newsletter »

Events

Linley Spring Processor Conference 2021
April 19 - 23, 2021
Virtual Event
More Events »