PIMIC Unveils Business Strategy, and Silicon Technology to Revolutionize AI at the Edge

Ultra-Efficient Silicon Solutions Deliver Industry-Leading Performance and Energy Savings for Next Generation Smart Devices

Cupertino, CA – December 11, 2024 – PIMIC, a groundbreaking AI semiconductor startup, today announced its official corporate launch, unveiling its mission to deliver cutting-edge AI inferencing silicon designed to power the next generation of intelligent, voice-activated edge devices (VAD) including toys, home and business audio systems, wearables, and robots.

PIMIC was founded by technology industry veteran Subi Krishnamurthy, CEO/CTO, who has assembled a world-class engineering and business development team designing and delivering highly-efficient edge AI solutions with unparalleled performance and energy savings. The company was founded in response to two significant market trends: the rapid shift of AI inference to edge devices, and the rise of speech as the primary interface for intelligent interactions.

"We founded PIMIC to empower AI devices at the edge with the kind of efficiency and performance that has, until now, only been possible in large, power-intensive environments," said Subi Krishnamurthy, CEO/CTO of PIMIC. "We recognized a need for AI to be more accessible and practical, not just for cutting-edge technology users, but for everyday applications that can enhance lives."

PIMIC’s proprietary Jetstreme™ Processing-in-Memory (PIM) acceleration architecture brings remarkable gains in AI computing efficiency by addressing the key requirements in edge environments, including low power, compact design, and superior AI model parameter update performance.

“Jetstreme provides substantial technological flexibility, allowing PIMIC to port across semiconductor manufacturing nodes with minimal incremental investment. With its pioneering memory structure, PIMIC aims to redefine the edge AI market by offering a solution that brings cutting-edge innovation to diverse industry segments,” continued Subi Krishnamurthy.

The Technology: A collaboration with leading semiconductor foundry partner

Jetstreme is currently built on TSMC’s 22nm ultra-low leakage (ULL) technology platform, offering ULL SRAM solutions that enables PIMIC’s scalable AI memory architecture.

Key advantages of Jetstream include:

• 10x Power, Performance, and Area (PPA) advantage over existing near-memory solutions
• Dynamic AI model updates with high write performance of model parameters, essential for next-generation edge devices
• Model reconfiguration consumes very low energy enabling prolonged operational life in battery operated edge AI applications

Jetstreme’s 10x PPA advantage over existing near-memory solutions transforms how intelligent devices operate and enables new possibilities in audio and other edge AI applications.

“AI is rapidly expanding from the cloud to the edge, driven by the growing demand for real-time processing, low latency, and energy-efficient solutions in edge devices,” said Lucas Tsai, Vice President of Business Management at TSMC North America. “We are proud to collaborate with innovators like PIMIC to tackle the challenge of efficient memory architectures and power the next generation of edge AI applications using TSMC’s industry-leading ultra-low-power platform.”

Addressing the Shift to Edge AI

According to global market intelligence, data and events provider IDC, the AI industry is witnessing a rapid shift towards the edge as demand grows for localized, low-latency, and cost-efficient AI solutions. IDC forecasts $41 billion in edge endpoint AI processor and accelerator revenue in 2028 - a little more than double that of 2023. (Source: IDC’s Worldwide Edge Endpoint AI Processor and Accelerator Forecast, 2024–2028, May 2024.) Market trends highlight the need for smaller, high-quality language models that can operate locally, reducing dependence on cloud infrastructure. PIMIC’s technology is designed to capitalize on these trends, demonstrating how AI-driven speech recognition, noise cancellation, and natural language processing can all be achieved on-device with minimal power requirements.

"Endpoint devices are increasingly adding on-device AI capabilities that require very small die size and very low power consumption,” said Phil Solis, Research Director at IDC. "PIMIC's solutions uniquely address this and are small enough to fit inside MEMS sensor devices, making them easy to integrate into existing solutions, and are also scalable enough to run AI models up to billions of parameters while eliminating the memory interface bottleneck."

Availability

Jetstreme product design and development services are available now by contacting: contact@pimic.ai. Products based on Jetstreme technology will be available in early 2026.

PIMIC will be at CES 2025, January 7-10, Las Vegas Convention Center, North Hall, Booth #9628

About PIMIC

Founded in 2022 and based in Cupertino, California, PIMIC is an AI semiconductor company specializing in ultra-efficient silicon solutions for edge AI applications. The company’s chip products deliver industry-leading performance and power efficiency, enabling advanced AI capabilities in compact, low-power devices. With a focus on empowering devices at the edge, PIMIC aims to redefine how AI is integrated into everyday technology.

For more information, visit www.pimic.ai.