ANAFLASH Advances Embedded FLASH Memory for Next-Generation Smart Edge Devices with Samsung Foundry
SUNNYVALE, Calif., November 3, 2025 – ANAFLASH, an edge AI processor startup, has developed an AI MCU with its proprietary standard logic compatible embedded flash technology using Samsung Foundry’s 28 nm process. This breakthrough enables the integration of zero-standby-power weight memory based on ANAFLASH’s Logic-EFLASH technology, utilizing only standard logic devices. The result is a cost-effective and energy-efficient AI MCU designed for battery-powered smart edge devices.
“Supporting startups like ANAFLASH reflects Samsung Electronics’ commitment to expanding the versatile silicon manufacturing ecosystem with innovative technologies,” said Margaret Han, Executive VP and Head of US Foundry at Samsung Electronics. “We look forward to continuing our collaboration with ANAFLASH to advance a cost-effective way to integrate the nonvolatile memory technology in our advanced logic processes, unlocking new possibilities of intelligent edge computing.”
ANAFLASH’s collaboration with Samsung Electronics originated through its wholly owned subsidiary in Korea, SEMIBRAIN, the inaugural winner of the fabless challenge contest that the company sponsored in 2022. Since then, multiple research grants were awarded by government funding agencies in Korea to further develop its Logic-EFLASH technology using Samsung Foundry and build a cost and energy efficient AI MCU based on the tightly coupled weight memory technology.
“The support from Samsung enabled us to prove our Logic-EFLASH technology for the first time using high-k and metal-gate technology," said Younghee Jeon, Head of Korea Research Center of ANAFLASH. “This achievement demonstrates how AI computation can be accelerated efficiently tightly coupled with the zero-standby power weight memory, and we will continue accelerating our innovation in the more advanced process nodes.”
The AI MCU built using a 28 nm process from Samsung Foundry also features a 4-bits/cell embedded flash technology demonstration to store a trained AI model efficiently. ANAFLASH will present its AI MCU architecture with the tightly coupled weight memory at the IEEE Asian Solid-State Circuits Conference, taking place on November 2-5, 2025 in Daejeon, Korea.
ANAFLASH’s AI MCU is currently sampled to the partners and customers with supporting software development packages.
Related Semiconductor IP
- 100-10Gbps Ethernet MAC Controller
- SD Card Controller
- AXI4 Memory Controller IP Core
- AXI-Lite I2C Controller Core
- LVDS Transceiver IP core
Related News
- GlobalFoundries and Microchip Announce Microchip's 28nm SuperFlash® Embedded Flash Memory Solution in Production
- Embedded Flash memory developer Floadia has raised 1.05 billion yen from INABATA and others
- Samsung Electronics and Broadcom Expand Strategic Collaboration Across Memory and Foundry Technologies
- RAAAM Memory Technologies Appoints Former Intel and Samsung Foundry Executive Hong Hao as Strategic Advisor
Latest News
- BrainChip Launches AKD1500 Raspberry Pi 5 Kit: Giving Developers a Direct Path from Prototype to Production
- Cadence Joins the Genesis Mission Consortium to Advance AI-Driven Scientific Discovery and Microelectronics Innovation
- Onsemi picks up Codasip’s low-end RISC-V business for $14.5 million
- Semidynamics and Nokia Explore Collaboration to Strengthen Europe’s Sovereign AI Capabilities
- VeriSilicon Introduces GLTiny™ 3D GPU Programming Model for Ultra-Low-Power, Memory-Efficient Wearable Devices