Rebellions Partners with Arm, Samsung Foundry, and ADTechnology on Next-Gen AI Computing Chiplet Technology
Seoul, October 15, 2024 -- Rebellions, a pioneering AI chip startup, today announced a collaboration with Arm, Samsung Foundry, and ADTechnology to develop an AI CPU chiplet platform. This initiative leverages advanced chiplet technology to address the growing demands of AI workloads in data centers and high-performance computing (HPC), promising unprecedented levels of performance and energy efficiency.
In partnership with these industry leaders, Rebellions aims to accelerate the development and deployment of its AI inference chiplet solution, optimized for power efficiency. The collaboration is expected to greatly enhance AI computing capabilities, setting a new standard for a scalable and efficient silicon solution that combines Rebellions’ expertise in chip design with extensive industry experience from Arm, Samsung Foundry, and ADTechnology.
Rebellions will integrate its AI accelerator, REBEL, with ADTechnology’s compute chiplet powered by Arm® Neoverse™ Compute Subsystems (CSS) V3 technology. ADTechnology will design and implement the solution using Samsung Foundry’s advanced 2nm cutting-edge process technology, ensuring the production of high-quality chips.
“Rebellions was one of the first AI chip startups to boldly embrace chiplet technology, recognizing its potential to meet the diverse needs of AI workloads, from large AI model builders to hyperscalers,” said Jinwook Oh, CTO of Rebellions. “We’re excited to lead this new phase of the chiplet ecosystem alongside our partners, leveraging our AI accelerator technology and experience to drive innovation in the era of generative AI.”
The collaboration builds on Rebellions’ state-of-the-art work with its REBEL chip, which is designed for scalability and equipped with high-bandwidth memory. Integrating REBEL into this AI CPU chiplet platform offers a future-proof acceleration solution for the ever-growing AI workload landscape, ensuring both energy efficiency and optimal performance with an estimated 2-3x efficiency advantage for GenAI workloads (Llama3.1 405B parameter LLMs).
“The rapid evolution of AI workloads demands specialized solutions that can deliver both performance and efficiency,” said Eddie Ramirez, vice president of go-to-market, Infrastructure Line of Business, Arm. “By tapping into the unique benefits enabled by the Arm Total Design ecosystem and flexibility of Arm CSS, this collaboration enables Rebellions to bring cutting-edge chiplets to market faster, which is critical to driving innovation in AI datacenters.”
“This multi-way collaboration leveraging Samsung Foundry’s state-of-the-art process technologies and advanced packaging solutions marks a significant milestone for the future of the AI semiconductor landscape, and will play a crucial role in building a robust AI semiconductor ecosystem in Korea,” said Taejoong Song, Vice President and Head of Foundry Business Development at Samsung Electronics.
About Rebellions
Rebellions is a pioneering AI chipmaker that specializes in chips for AI inference, bringing superior energy efficiency as well as performance. Founded in 2020, the company has introduced two groundbreaking chips within just three years of its inception: the finance market-focused ION and the datacenter-focused ATOM. ATOM has been successfully deployed in a large-scale datacenter run by KT(Korea Telecom). Currently, Rebellions is developing its next-generation AI chip, REBEL, featuring a scalable chiplet architecture and high-bandwidth 144GB HBM3e memory. which is engineered to enhance Large Language Models (LLMs) and Multi-Modal Models (MMMs).
For more information: https://rebellions.ai
Related Chiplet
- Direct Chiplet Interface
- HBM3e Advanced-packaging chiplet for all workloads
- UCIe AP based 8-bit 170-Gsps Chiplet Transceiver
- UCIe based 8-bit 48-Gsps Transceiver
- UCIe based 12-bit 12-Gsps Transceiver
Related News
- ADTechnology Partners with Arm, Samsung Foundry, and Rebellions on AI CPU Chiplet Platform
- Samsung and Arm Forge Ahead with Next-Gen AI Chiplet Technology
- Eliyan Ports Industry’s Highest Performing PHY to Samsung Foundry SF4X Process Node, Achieving up to 40 Gbps Bandwidth at Unprecedented Power Levels with UCIe-Compliant Chiplet Interconnect Technology
- Tenstorrent Selects Samsung Foundry to Manufacture Next-Generation AI Chiplet
Latest News
- Alphawave Semi Scales UCIe™ to 64 Gbps Enabling >20 Tbps/mm Bandwidth Density for Die-to-Die Chiplet Connectivity
- Silicon Box welcomes European Commission approval of €1.3 billion Italian State aid measure to support new advanced packaging facility in Novara
- Fraunhofer IMS Takes a Key Role in Establishing the APECS Pilot Line
- EdgeCortix Joins AI-RAN Alliance to Accelerate the Integration of AI and Next-gen RAN Infrastructure
- Cadence Rolls Out System Chiplet to Reorganize the SoC