en.Wedoany.com Reported - YiXing Intelligence (Beijing) Technology Co., Ltd., a Chinese RISC-V cloud AI computing power chip company, recently signed a strategic cooperation agreement with China Energy Engineering Corporation (Beijing) Smart Computing Power Technology Co., Ltd. and Beijing Guofu Cloud Data Technology Co., Ltd. The three parties will collaborate on the construction and operation of computing power centers, the AI industry ecosystem, and the application deployment of AI computing power in China. Together with ecosystem partners, they aim to build the world's first full-stack RISC-V supernode AI Token factory.
This project integrates RISC-V chips, supernode computing architecture, computing power center operations, AI application services, and the industry ecosystem into a unified computing system, providing Token generation capabilities for large model inference and agent services.
Tokens are becoming the fundamental unit for AI service costs and output. When enterprises use large models for Q&A, code generation, knowledge base retrieval, intelligent customer service, office automation, and agent tasks, tokens are continuously consumed behind the scenes. The energy consumption, cost, throughput, and stability per token directly impact the commercialization potential of AI service platforms. The collaboration between YiXing Intelligence, China Energy Engineering Corporation Smart Computing Power, and Guofu Data addresses the issues of computing power supply, energy coordination, and operational efficiency following the large-scale application of large model services. For enterprise customers, whether a computing power center can stably output tokens is closer to real usage needs than simply comparing peak chip parameters.
The RISC-V architecture is a technical feature of this project. YiXing Intelligence adopts a RISC-V approach combined with its self-developed virtual instruction set, building a Chinese AI computing power foundation based on an open instruction set.
Public information shows that YiXing Intelligence has launched the Epoch series of AI computing chips and related computing platform products. The Epoch, built on the RISC-V instruction set and the self-developed VISA virtual instruction set, has entered the stage of mass production and has formed a product layout covering chips, computing modules, complete machines, and large-scale system solutions. The AI Token factory needs to verify not only the performance of individual chips but also the continuous operation capability of chip clusters within the software stack, scheduling system, network interconnection, cooling and power supply, and operation and maintenance systems. When facing high concurrency requests, large model inference and agent services continuously test inter-chip communication, memory bandwidth, task scheduling, system fault tolerance, and resource utilization efficiency. The supernode integrates multiple AI chips, multiple servers, interconnection networks, storage, and software scheduling into a larger logical computing unit, transforming the capabilities of individual chips into system capabilities that can provide external services.
The involvement of China Energy Engineering Corporation Smart Computing Power enables the project to coordinate computing power center construction with energy infrastructure. The AI Token factory requires capabilities in data center rooms, power supply, liquid cooling or efficient heat dissipation, rack deployment, operation and maintenance systems, and energy management.
Guofu Data complements the project with data center and cloud data operation capabilities. An AI Token factory serving enterprise customers and industry applications requires network access, resource scheduling, security compliance, multi-tenant management, data governance, and customer service systems. After the tripartite collaboration, RISC-V computing power can move from the chip and server stage into a service system oriented towards external customers, providing a more complete AI computing power supply solution for model companies, agent companies, and government and enterprise clients. The initial supernode deployment scale, supported model types, energy consumption per token, inference throughput, operational stability, and customer usage will determine the demonstrative value of this project.
The cooperation between YiXing Intelligence, China Energy Engineering Corporation Smart Computing Power, and Guofu Data pushes China's RISC-V AI chips closer to system scenarios for commercial operation. If real-world business verification proceeds smoothly, RISC-V AI computing power will gain a clearer engineering path.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
