Rebellions selects arteris for its next-generation neural processing unit aimed at generative AI

Arteris provider of system IP which accelerates system-on-chip (SoC) creation, announced that Rebellions, a pioneering AI semiconductor startup, will deploy FlexNoC interconnect IP plus Magillem Connectivity and Magillem Registers for its next-generation AI Hardware Accelerator Neural Processing Unit. The combination of Rebellions AI technology with Arteris’ system IP will provide maximum flexibility and energy efficiency, and enable companies to create powerful, cost-efficient AI hardware at scale.

Rebellions’ products increase data execution with minimum power consumption. They are addressing the needs of latency-critical AI inference applications from Generative AI and Large Language Models, supporting finance and cloud computing industries with their cutting-edge chips. Rebellions is also developing its next-generation chip in collaboration with Samsung.

“The market for AI, including Generative AI applications, is exploding and optimized data flow and execution is paramount for complex computations,” said K. Charles Janac, president and CEO of Arteris. “Our silicon-proven IP coupled with our SoC integration automation software will support Rebellions in delivering high-performing, energy-efficient AI products with aggressive time-to-market targets. We are very pleased to work with such a visionary AI startup.“

Arteris’ FlexNoC network-on-chip interconnect IP provides scalability, low latency and power-efficient on-chip communication for superior performance. Arteris’ Magillem Connectivity combined with Magillem Registers software automates semiconductor design flows, boosting productivity and improving quality with early error detection. Arteris system IP products enable fast, flexible and energy-efficient SoC creation, meeting the performance and schedule needs in the most demanding markets.