AnalogAI's edge AI processors incorporate a proprietary hardware-aware algorithm that has been designed to simultaneously train and run inference on AI models in real-world environments.

AnalogAI uses SST's memBrain SAGE IP to deliver high levels of aCIM (analog compute-in-memory) performance with power consumption on the order of one watt or less in ultra-low-power edge applications. Thanks to SST's powerful, silicon-proven, SuperFlash®-based memBrain SAGE IP, AnalogAI is developing new capabilities for real-world inference applications such as humanoid robots, drones, and adaptive vehicles.

“AnalogAI develops innovative and optimized solutions for edge AI inference that enable devices to adapt in real time to sudden changes in the environment,” said Mark Reiten, vice president of Microchip’s Intelligent Computing business unit. “As the core of the inference engine for AnalogAI’s first products, SST’s memBrain SAGE IP provides the necessary computing performance along with the efficient power consumption that AnalogAI demands to meet its application goals. We are pleased to welcome AnalogAI as a new licensee of our SST memBrain IP, adding to our rapidly expanding ecosystem.”.

Features of the SST memBrain SAGE IP:
● memBrain TILE (Tensor In-Memory Logic Element)
or Optimized memBrain solution with ESF3 bit cell that stores up to 8 bpc with nanoamps
or memBrain solution with custom matrix, decoders and driver circuitry
● Optimized DAC and ADC

● High-voltage adder and bias circuitry
● Nanoampere bit cell control logic
● Proprietary memBrain test circuitry
● SST memBrain technology documentation and simulation modes
● Full IP integration service and support

“AnalogAI selected SST’s memBrain SAGE IP after evaluating the available options on the market and concluded that the memBrain SAGE IP is the best for accelerating our development timeline while maintaining the extremely low power consumption and high performance demanded by our market segment,” said Jaejun Lee, CEO of AnalogAI. “aCIM is a rapidly emerging field, and at AnalogAI, we are at the forefront of the new functionality that edge AI devices must offer to adapt to the real world.”.

SST 's memBrain SAGE IP has been developed and implemented on 40nm and 28nm manufacturing processes with production-ready SuperFlash memory. Current technology development plans include the development of the 22nm memBrain SAGE IP. SuperFlash memory, designed to provide reliable, high-performance, low-power on-chip non-volatile storage, is widely used in applications that demand fast access times, long data retention, and long lifespan without the need for external memory components.