Zhongke Tongliang Showcases Full-Stack RISC-V Video and AI Computing Products at Shenzhen Summit

The summit was hosted by the RISC-V International Open Source Laboratory, also known as RIOS Lab. Publicly available information cited in the report says RISC-V now accounts for 25% of the global processor market, meaning one in four processors is based on the architecture. The report describes the summit as a window on the evolution of the global RISC-V ecosystem and industrial deployment, as the architecture moves from embedded and IoT markets toward high-performance computing, data centers and AI video.

Zhongke Tongliang says its core technical team has more than 20 years of experience in video intelligence and has won first place in both the large and small dataset categories of GreenGraph 500. The company says it holds more than 140 granted patents and that R&D staff account for more than 50% of its workforce. Its core technical path combines a dataflow-native architecture with the RISC-V open ecosystem, with the aim of redefining the computing model for video intelligence.

The SmarCo AI workstation is positioned as a desktop-level high-performance intelligent computing terminal for video intelligence, AI large models, robot debugging and other scenarios. According to the report, it features a 12-core RISC-V processor running at 2.0GHz, 250 TOPS INT8 computing power, 128GB of LPDDR5 ECC unified memory, support for real-time hardware decoding of 128 channels of 1080P video and fully open interfaces. Developers at the booth asked about the RISC-V core architecture, memory configuration, the degree of interface openness and adaptation for large-model inference, video processing and robotics, the report said.

The SmarCo intelligent video encoding appliance centers on chip-level 'lossless' intelligent encoding and targets the transmission and storage of large volumes of video. The company demonstrated the product on site and highlighted features including reuse of existing capacity with migration in 30 to 90 days, access without changing the original architecture, data security protection and self-controlled core technology. Visitors from video storage, security and industry informatization asked about encoding efficiency, storage expansion methods, compatibility with existing architectures and data security mechanisms, according to the report.

The SmarCo intelligent computing accelerator card targets video and AI computing scenarios and uses the RISC-V open ecosystem as its foundation. The report said the product is intended to provide higher energy efficiency and lower-threshold intelligent computing acceleration. Visitors asked about application scenarios, adaptation plans and performance, while some ecosystem partners held preliminary discussions with staff on joint solutions and software toolchain adaptation.

The booth drew developers who have long followed RISC-V as well as industry customers from video, security, intelligent computing, robotics and industrial sectors, along with investors and ecosystem partners. Questions covered RISC-V performance in video encoding and decoding, the meaning of fully open interfaces for research and debugging, how the video encoding appliance avoids changing the original architecture and whether the accelerator card can adapt to existing AI frameworks. Some visitors left contact details for product trials, ecosystem adaptation, joint solutions and commercial cooperation.

A Zhongke Tongliang representative at the event said the value of RISC-V depends on deep integration with specific application scenarios and on forming a mass-producible, replicable commercial loop, rather than on the instruction set alone. The representative said video is one of the scenarios with the largest data volumes, the highest real-time requirements and the broadest applications, while intelligent computing is core infrastructure in the AI era. The company hopes to combine its dataflow-native architecture with the RISC-V open ecosystem to explore a new computing model for video intelligence that retains the flexibility and self-control of an open architecture while optimizing for video and intelligent computing scenarios.

Editor's Summary

Zhongke Tongliang used the Sixth RISC-V China Summit to present a full-stack RISC-V product line spanning AI workstations, video encoding appliances and intelligent computing accelerator cards. The products target video and AI computing scenarios, where the company says open RISC-V architecture can be combined with dataflow-native design for commercial deployment. The event also served as a venue for developer and partner discussions on adaptation, trials and joint solutions.