日新技研株式会社
日新技研株式会社
  • ホーム
  • 日本語
  • English
Architecting the Future of Intelligent Interconnect

Welcome to NISSHINGIKEN!

We're a deep-tech interconnect IP company based in Kyoto, Japan — with teams in US and Hong Kong. We build the silicon infrastructure that powers the next generation of AI, HPC, and data center systems.

Contact Us

What We Do

High-Performance Memory Subsystem IP

High-Performance Memory Subsystem IP

High-Performance Memory Subsystem IP

We deliver high-performance memory subsystem IP, including NoC, advanced cache architectures, and high-bandwidth memory controllers. Our solutions are optimized for data-intensive workloads, enabling scalable and efficient infrastructure for next-generation AI and data center systems.

Chiplet IO Die Design

High-Performance Memory Subsystem IP

High-Performance Memory Subsystem IP

Leveraging UCIe technology, we specialize in high-speed Chiplet IO Die design. Our solutions minimize P2P latency and maximize interconnect bandwidth — enabling scalable, efficient heterogeneous computing for next-generation AI accelerators and scale-up networks.

SoC Architecture Consulting

High-Performance Memory Subsystem IP

SoC Architecture Consulting

From workload-driven performance analysis to RTL-level micro-architecture definition, we deliver end-to-end SoC architecture consulting. We optimize NoC topology, system efficiency, and high-performance scalability — accelerating time-to-market and reducing design risk.

Our Products

High-Performance NoC Interconnect

Highly configurable Mesh, Ring, and Crossbar topologies with full AXI4 / AHB / APB compliance. Featuring advanced QoS management and Virtual Channel flow control to eliminate data congestion in complex SoCs.

Scalable System Cache

Optimized for near-memory computing with fully parameterizable ways and sets.
With integrated ECC protection and co-design with our NoC IP, the architecture enables low-latency, high-efficiency data access.

Next-Gen UCIe Die-to-Die Interconnect

Fully compliant with UCIe 1.1 / 2.0 standards, featuring a robust AXI4 transaction layer and protocol flit bridging for PCIe, CXL, and streaming protocols — enabling efficient chiplet integration across heterogeneous architectures.

112G Ultra-High-Speed Ethernet

Advanced 112G Serdes PCS supporting SUE / UEC / UALink protocols. Designed for data center scale, featuring link-layer retry, CBFC (Credit-Based Flow Control), and  supports integration with major switch vendors through configurable behavior. 

Contact Us

We welcome partnerships and inquiries from teams working on next-generation AI, HPC, and data center silicon.

NISSHINGIKEN Co., Ltd.

KRP Building No.4 3F, 93 Chūdōji Awatachō, Shimogyo Ward, Kyoto, 600-8815, Japan

075-315-8861 info@nisshin-giken.jp

Business Hours

Today

Closed

Schedule a Discussion

Attach Files
Attachments (0)

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

Cancel

日新技研株式会社

Copyright © 2026 日新技研株式会社 - All Rights Reserved.

  • 日本語
  • English

Powered by

This website uses cookies.

We use cookies to analyze website traffic and optimize your website experience. By accepting our use of cookies, your data will be aggregated with all other user data.

Accept