PIMIC Presents Business Strategy and Silicon Technology to Advance AI at the Edge – Embedded

PIMIC, an innovative AI semiconductor startup, has officially launched, revealing its objective to provide advanced AI inferencing silicon intended to enhance the forthcoming generation of intelligent, voice-activated edge devices (VAD), such as toys, audio systems for homes and businesses, wearables, and robots.
PIMIC was established by technology industry expert Subi Krishnamurthy, CEO/CTO, who has curated a premier engineering and business development team focused on building and delivering highly efficient edge AI solutions with exceptional performance and energy conservation. The company was established in reaction to two major market trends: the swift transition of AI inference to edge devices and the emergence of speech as the principal interface for intelligent interactions.
PIMIC’s unique Jetstreme™ Processing-in-Memory (PIM) acceleration architecture significantly enhances AI processing efficiency by meeting essential criteria in edge contexts, such as low power consumption, small design, and exceptional speed in updating AI model parameters.
Jetstreme is presently constructed on TSMC’s 22nm ultra-low leakage (ULL) technology platform, providing ULL SRAM solutions that facilitate PIMIC’s scalable AI memory architecture. Jetstreme’s 10x PPA superiority over current near-memory solutions revolutionizes the functionality of intelligent devices and facilitates novel opportunities in audio and other edge AI applications.
Global market intelligence company IDC reports that the AI sector is experiencing a swift transition towards edge computing due to increasing demand for localized, low-latency, and cost-effective AI solutions. IDC projects $41 billion in sales from edge endpoint AI processors and accelerators by 2028, which is slightly more than double the amount anticipated for 2023 (Source: IDC’s Worldwide Edge Endpoint AI Processor and Accelerator Forecast, 2024–2028, May 2024). Market trends underscore the necessity for compact, high-caliber language models capable of local operation, hence diminishing reliance on cloud infrastructure. PIMIC’s solution aims to leverage these developments, illustrating that AI-driven speech recognition, noise suppression, and natural language processing can be executed on-device with minimum power consumption.

source
This article was autogenerated from a news feed from CDO TIMES selected high quality news and research sources. There was no editorial review conducted beyond that by CDO TIMES staff. Need help with any of the topics in our articles? Schedule your free CDO TIMES Tech Navigator call today to stay ahead of the curve and gain insider advantages to propel your business!