Artificial Intelligence

PFN and SEMIFIVE Collaborate to Develop MN-Core L2000 AI Inference Chips

· GlobeWire Daily Editorial Desk · Source: www.prnewswire.com

2 min read

Key takeawaysPreferred Networks, Inc. (PFN) and SEMIFIVE Inc. have announced a collaboration to develop the MN-Core L2000 Series, a next-generation line of AI inference chips designed to succeed the MN-Core L1000 Series. The announcement states that the new chip series aims to deliver more than twice the memory bandwidth of its predecessor.
PFN and SEMIFIVE Collaborate to Develop the MN-Core L2000 Series of Next-Generation AI Inference Chips

Image: SEMIFIVE · Photo source · Resized and letterboxed to 16:9; no AI generation

Preferred Networks, Inc. (PFN) and SEMIFIVE Inc. have announced a collaboration to develop the MN-Core L2000 Series, a next-generation line of AI inference chips designed to succeed the MN-Core L1000 Series. The announcement states that the new chip series aims to deliver more than twice the memory bandwidth of its predecessor.

According to the companies, the collaboration is underway with a tape-out scheduled for the first quarter of 2027. The project utilizes a 3D-stacked structure that vertically integrates logic chips with dynamic random access memory (DRAM) to address bandwidth limitations in generative AI workloads.

Key Details of the Collaboration

  • Roles: PFN is managing processor architecture and logical design, while SEMIFIVE is providing turnkey design implementation and supply chain management.
  • Manufacturing: The logic chips are planned to be manufactured using Samsung Foundry's platform technology and 3D-IC design infrastructure.
  • Exhibition: The companies plan to jointly exhibit the MN-Core L2000 Series at Booth 4606 during the SC26 event in Chicago.
  • Funding Support: The development is supported by Japan's Acquisition, Technology & Logistics Agency (ATLA) under a specific research project running from fiscal year 2024 to 2028.

Technical Approach and Goals

The MN-Core L Series focuses on memory bandwidth as a key factor for generative AI inference performance. By combining high computational performance with low power consumption in a 3D-stacked architecture, the initiative seeks to improve power efficiency and processing speed for AI agents and infrastructure applications.

Readers and industry observers can verify the original announcement through official press release distributions and company announcements issued by PFN and SEMIFIVE.

Share this article
Read the original source: PFN and SEMIFIVE Collaborate to Develop the MN-Core L2000 Series of Next-Generation AI Inference Chips

AI-assisted reporting; source claims are attributed and have not been independently verified.

About the author

GlobeWire Daily Editorial Desk — AI-assisted news based on attributed published sources.

More to read