NuAxi
NuAxiSemiconductor trade intelligence

What If Your Memory Could Think?

What happened

Per the September 2026 announcement, AnalogAI, a South Korean company specializing in edge AI processors, has selected memBrain SAGE (Synaptic Analog Generative Engine) neuromorphic hardware IP from Silicon Storage Technology (SST), a Microchip subsidiary, for its first real-world edge AI processors, targeting inference with on-device adaptation to changing real-world conditions at a power consumption of one watt or less; the article notes it is early days for AnalogAI's processors, and the source does not state the cause of any price or supply impact. AnalogAI is SST's fourth publicly announced memBrain licensee, following Witmem in February 2022 (low-power speech processing), Intelligent Hardware Korea in September 2023 (analog compute platforms for edge AI, particularly computer vision), the October 2024 announcement by GlobalFoundries, Witmem and Microchip of sampling of Witmem's WTM-8 video processor capable of 8K video at 120 frames per second, and Mythic's March 2026 selection of memBrain technology for its next generation of ultra-low-power analog processing units. The technology is built on SST's embedded SuperFlash memory using established 40nm and 28nm foundry processes; in SST's YOLOv5 comparison example, the aCIM implementation processes 150 frames per second at 3 watts versus the GPU-based implementation's 200 frames per second at 60 watts, or roughly 15 times the energy efficiency (the THUS AI audio chip in Anker's Soundcore Liberty 5 Pro earbuds also uses this technology, with underlying silicon supplied by Witmem).

Impact and signal

Low impactImpact score: 10/100

Affected components and suppliers

flash memoryedge AI processors
AnalogAISilicon Storage TechnologyMicrochip Technology
memBrain SAGE IPanalog compute-in-memoryedge AIIP licensingultra-low-power

Source and time

Source: EE Journal · 10/27/2026 · View original source

Open the NuAxi live feed →

Real-time pricing, supply signals and AI analysis for semiconductor traders