New collaboration delivers full-stack computing solutions that can support large-scale AI advancements in the public and private sectors. $18.6 million Grant

Santa Clara, California, March 6, 2024 /PRNewswire/ — Today, Encharge AIAnnounces partnership with company commercializing next-generation AI accelerators princeton universityis being supported by the U.S. Defense Advanced Research Projects Agency (DARPA) to develop advanced processors that can run AI models more efficiently than previously thought possible.

DARPA's Optimum Processing Technology Inside Memory Arrays (OPTIMA) program $78 million An effort to develop fast, power-efficient, and scalable compute-in-memory accelerators that unlock new possibilities for commercial and defense AI workloads not possible with current technology. As part of OPTIMA, DARPA will $18.6 million grants for multi-year projects proposed by princeton university and EnCharge AI.

The rapid development of AI has rapidly increased processing requirements, which are now met by large server farms with high cost and power requirements. Extending AI adoption from the cloud to real-time, mission-critical applications at the edge requires a move beyond existing processing technologies.

DARPA's OPTIMA recognized the need for a major breakthrough rather than an optimization of existing GPUs and other digital accelerators. DARPA uses existing VLSI manufacturing technology and specifically excludes research that “primarily results in evolutionary improvements to existing practices,” while “innovative research that enables revolutionary advances in science, devices, and systems.” It called for funding a “new approach”.

The key to this is princeton university-EnCharge AI project is directed by Dr. Naveen VermaProfessor of Electrical and Computer Engineering princeton Co-founder and CEO of EnCharge AI. Many of the innovations that the OPTIMA project aims to further develop were discovered in Dr. Verma's research. princeton The lab was established in part with previous funding from DARPA and the Department of Defense.

This project uses next-generation switched-capacitor analog in-memory computing chips developed by Dr. Verma's lab and commercialized by EnCharge AI to advance AI applications and end-to-end workload execution. I'll investigate. “The future will be about decentralizing AI inference, taking it out of the data center and into phones, laptops, cars and factories,” he said. “EnCharge is currently bringing the first generation of this technology to market, but we look forward to supporting DARPA in accelerating the next generation and seeing how far we can improve performance and efficiency.” .”

Switched capacitor analog in-memory computing technology has been demonstrated in several generations of silicon developed in the United States. princeton It can improve computing efficiency by orders of magnitude compared to digital accelerators while maintaining accuracy and scalability not possible with current-based analog computing approaches. For more information about switched capacitor in-memory computing, please visit: A series of peer-reviewed foundational published papers. Co-authored with Dr. Verma.

EnCharge AI CTO Dr. Kailash Gopalakrishnan EnCharge AI noted that the constraints posed by current AI processing technology motivated it to join OPTIMA. “EnCharge brings together leaders from all over the world. princeton, IBM, Nvidia, Intel, AMD, Meta, Google, and other companies that have brought computing into the modern era. At some point, many of us realized that innovation within existing computing architectures and the scaling of silicon technology nodes is slowing down, just when AI was creating huge new demands for computing power and efficiency. I realized that. GPUs are the best tools available today, and we're excited that DARPA is supporting the development of a new generation of chips that will unlock the potential of AI. ”

Joining OPTIMA comes on the heels of the recently announced EnCharge AI. $22.6 million funding The round includes new investors VentureTech Alliance, RTX Ventures, and ACVC Partners to develop next-generation full-stack AI computing solutions that fundamentally transform the way AI is used from the edge to the cloud. Masu.

About EnCharge AI
EnCharge AI is a leader in advanced AI solutions for edge-to-cloud deployments. EnCharge's robust, scalable, next-generation in-memory computing technology provides orders of magnitude higher computing efficiency and density compared to today's best-in-class solutions. High-performance solutions enable the immense potential of AI to be harnessed at scale in power-, size-, and weight-constrained applications. Founded in 2022, EnCharge AI is led by veteran engineers with backgrounds in semiconductor design and AI systems. Raised by EnCharge AI $45 million up to now. For more information about EnCharge AI, please visit:

Media contact:
Sparkpr for EnCharge AI
[email protected]

Source EnCharge AI

Source link