With the recent emergence of energy consumption issues, power- and cost-efficient solutions using embedded flash (eFlash) memory are gaining attention.

ANAFLASH, a Silicon Valley-based developer of smart microcontrollers (MCUs) for edge computing, announced today that it has obtained an exclusive license for single-poly-based embedded flash memory technology from the University of Minnesota.

It is expected that the company will commercialize cost-effective and energy-efficient non-volatile memory technology and provide licensing for various battery-powered applications such as medical wearables, wireless sensors, and autonomous robots.

Energy-saving devices require embedded non-volatile memory to store information without consuming power in sleep mode. Unlike other types of alternative non-volatile memory technologies, AnaFlash eFlash memory technology is fully logic compatible and has the advantage of being scalable across advanced logic process technologies from a variety of silicon foundries. This explains that it can also alleviate cost and supply issues of semiconductor devices.

AnaFlash has proven this technology across multiple generations of silicon wafers and across a variety of process technology nodes, and has also demonstrated a prototype chip for a single-poly eFlash-based non-volatile neural network accelerator at Flash Memory Summit.

“We are excited about the opportunity in single-poly eFlash memory technology that can be easily deployed at advanced nodes without additional cost, as it does not require process overhead beyond standard logic processes,” said Song Seung-hwan, CEO of AnaFlash.

“The Berkeley SkyDeck Chip Track demonstrates that ANAFLASH’s licensed technology promises more efficient and lower-cost devices for a wide range of applications,” said Chon Tang, Founding Partner of the Berkeley SkyDeck Fund. “We look forward to continuing to leverage this important research to see how it can make a real difference.”

AnaFlash’s research partner, the Korea Advanced Institute of Science and Technology (KAIST), said it has built an energy-efficient edge AI test chip using the company’s logic-compatible eFlash IP, which will be presented at the 2023 IEEE Custom Integrated Circuits Conference in San Antonio, Texas, April 23-26.

Meanwhile, AnaFlash has received several awards from the National Science Foundation, the Department of Defense, and NASA, and recently closed a funding round led by Mirae Asset Venture Investment and We Ventures.

It also holds several patents in the United States and other countries, and has also established a wholly owned subsidiary in Korea called SEMIBRAIN. Last year, Semibrain was given the opportunity to develop embedded flash memory IP on Samsung Foundry’s advanced process node.

The single-poly-based eFlash memory technology was developed by Professor Chris Kim of the University of Minnesota and won the Low Power Design Contest award from IEEE. Professor Kim’s team has continued to develop this technology for various applications such as embedded nonvolatile memory, counterfeit detection sensors, and neuromorphic computing cores, and has published more than 10 papers.