en.Wedoany.com Reported - CCS-1 series, the first edge-side productivity AI chip independently developed by Chinese edge-side AI inference company Yuanli Semiconductor, has recently completed tape-out and been successfully powered on—achieving first-pass tape-out success, first-pass power-on success, and full compliance with all specifications. The tape-out and power-on of the first chip represents a comprehensive test of the team's system capabilities, marking the verification of the core technical roadmap and engineering capabilities, a critical step forward in the productization process, and injecting certainty into subsequent mass production and scaled delivery.
The success of the CCS-1 series did not begin on the tape-out floor, but rather stems from a complete technical system accumulated by Yuanli Semiconductor over a long period. Since its establishment in April 2023, the company has been building its product roadmap around the Chiplet modular design concept. Through its self-developed multimodal computing core and adaptive computing fusion technology, it binds the specification flexibility and cost constraints of AI inference chips within the same architectural framework. The interconnection efficiency between chiplets is the core variable of this technical route. Yuanli Semiconductor has introduced a self-developed interconnect architecture in the chip to work with the UCIe standard, ensuring that different computing modules collaborate with the lowest power consumption and highest bandwidth.
At the core performance level, the CCS-1 series has already demonstrated significant measured advantages over international flagship competitors, with relevant details to be fully disclosed during the subsequent product launch phase. The chip adopts a building-block architecture design, supporting enterprise customers in flexibly configuring computing specifications according to different scenario requirements, covering tasks such as large model local inference and edge-side training fine-tuning, while delivering high-performance edge-side inference capabilities under the triple constraints of power consumption, cost, and physical volume.
Yuanli Semiconductor is underpinned by a highly mature technical team. Founder and CEO Dr. Fang Shaoxia graduated from the Department of Electronic Engineering at Tsinghua University and has previously served as Global Chip R&D Director at AMD, AI Processor R&D Director at Xilinx, and Chip R&D Director and Chief Architect at DeePhi Tech. He has deep cultivation in the field of high-performance processors and AI chip architecture for over a decade, holding more than 50 AI chip-related invention patents in China and the US. The core team as a whole comes from international semiconductor companies such as AMD and Xilinx, having experienced the complete R&D and engineering cycles of over a dozen generations of AI chip architectures, and possessing full-chain closed-loop experience from architecture definition to large-scale mass production. The company has assembled a complete engineering team covering architecture design, chip implementation, system software, and business development, supporting the full process advancement from tape-out to delivery.
Strategic reserves on the capital side provide flexible space for the mass production advancement of CCS-1. On April 29, 2026, Yuanli Semiconductor completed a Pre-A round of financing exceeding 500 million yuan, led by IDG Capital, with participation from institutions such as Wuyuefeng SciTech and China Reform Fund, and continued increased investment from new and existing shareholders including Shangshi Capital, InnoAngel Fund, and CAS Star. This round of funding is explicitly directed towards the R&D of next-generation edge-side inference chips and the mass production advancement of a series of computing products, precisely covering the critical window for the CCS-1 series from tape-out verification to initial delivery.
The smooth power-on of the CCS-1 series is not just a technical delivery of a chip for Yuanli Semiconductor, but also a confirmation of strategic rhythm. The current edge-side AI inference market is on the eve of an explosion, with the demand for edge-side deployment of hundred-billion-parameter large models continuously expanding, while chip solutions that can achieve a practical balance between power consumption, cost, and computing power remain scarce. Yuanli Semiconductor enters this track with a Chiplet architecture, shortening the cycle verification from design to delivery with a single successful tape-out, thereby buying a time window for forming scaled deployments in subsequent scenarios such as PCs, edge servers, and industrial intelligent terminals.
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
