China’s RISC-V push hits a new milestone

China just put two more markers on the RISC-V map. At the 2026 Zhongguancun Forum, the Chinese Academy of Sciences said its open-source processor system Xiangshan set a new international record, while its native RISC-V operating system Ruyi became the first to support the RVA23 high-performance international standard.

That matters because RISC-V is no longer just a chip-nerd curiosity. It is becoming a serious option for countries and companies that want more control over processor design, software stacks, and long-term supply chains. China’s latest announcement shows it is trying to build more than chips. It is building a full stack around them.

What CAS actually announced

The headline from Beijing is simple: CAS says it has built a full-chain RISC-V development system that covers innovation, application, and ecosystem building. In plain English, that means the work is not limited to a processor core in a lab. It includes software, IP blocks, reference designs, and industrial partnerships that can move from research into products.

The two projects CAS highlighted are doing different jobs. Xiangshan is the high-performance processor side of the story. Ruyi is the software side, and its support for RVA23 gives it a standards-based anchor that matters for developers and vendors who want compatibility, not just a local demo.

CAS vice president Ding Chibiao said the academy has built a “Chinese solution” for the country’s information technology industry. That phrase is doing a lot of work, but the underlying message is clear: China wants a domestic RISC-V stack that can compete with proprietary architecture ecosystems.

• Xiangshan is described as an open-source high-performance RISC-V processor system.
• CAS says it includes the world’s first open-source network-on-chip IP.
• Ruyi is the first native RISC-V operating system to support RVA23.
• CAS says dozens of companies joined related development work.
• ZTE, Alibaba, and Tencent were named among the collaborators.

Why Xiangshan and Ruyi matter together

RISC-V has always had a clean pitch: open instruction set, low friction, and room for custom designs. The hard part has been everything around the core. If you want real adoption, you need toolchains, operating systems, interconnects, and actual silicon that can ship. That is where Xiangshan and Ruyi become more interesting than a press release would suggest.

Xiangshan gives the hardware story a stronger benchmark. CAS said it set a new international record, though the announcement did not spell out the exact metric in the English-language report. The mention of the world’s first open-source network-on-chip IP is more concrete. Network-on-chip IP is one of those quiet pieces of chip design that rarely gets hype, yet it matters a lot for how data moves inside complex processors.

“The open-source software movement and the open-source hardware movement are both about taking control of the tools we use,” said Linus Torvalds in a 2016 interview with the Linux Foundation.

That quote fits the moment well. RISC-V is following the same logic that made Linux influential: if enough of the stack is open, more people can build on it, audit it, and adapt it without waiting for a single vendor’s blessing.

Ruyi matters for the same reason. An operating system is where hardware becomes useful to real developers. If Ruyi can support RVA23 early, it gives chip designers and software teams a cleaner target for compatibility work. That is especially important in a market where software support often decides whether a chip becomes a product or stays a prototype.

The industrial list is the real signal

CAS said it worked with dozens of leading companies, including ZTE, Alibaba, and Tencent, to develop key technologies. That detail is more important than the ceremony around the forum. When large companies begin building on the same architecture, the ecosystem gets harder to ignore.

It also signals that the work is moving beyond academic proof-of-concept. Several companies have already built their own chips based on Xiangshan, which suggests the design is usable enough for follow-on work. In chip development, that is a meaningful threshold. A processor architecture can be open-source and still fail if nobody can turn it into something practical.

China is also trying to answer a strategic question that keeps coming up in semiconductors: how do you reduce dependence on closed ecosystems without sacrificing performance? RISC-V is attractive because it lets companies design around specific needs instead of paying for a one-size-fits-all license model.

• Arm licenses dominate much of today’s mobile and embedded chip market.
• RISC-V uses an open instruction set, which lowers entry barriers for custom silicon work.
• China’s announcement ties hardware, software, and industrial partners into one program.
• That mix is what turns an architecture from a research topic into a supply-chain option.

How this compares with the broader RISC-V race

China is not alone in betting on RISC-V. The architecture has been gaining traction across universities, startups, and major chip teams in the US, Europe, and Asia. But the scale of the CAS announcement is notable because it combines a processor system, an operating system, and industry participation in one public push.

For comparison, much of the global RISC-V story has centered on individual cores, development boards, or narrow embedded use cases. Those are useful, but they do not amount to a full ecosystem. CAS is trying to show breadth. The organization is talking about open-source IP, standards support, and commercial follow-through in the same breath.

That approach could matter for China’s domestic chip market in a practical way. If more companies can reuse shared RISC-V building blocks, they may move faster on specialized chips for networking, AI edge devices, and industrial systems. The payoff is less about one perfect processor and more about reducing duplication across many designs.

• Arm’s model depends on licensing; RISC-V does not.
• Many RISC-V efforts begin with cores; CAS is pairing cores with an OS and IP blocks.
• RVA23 support gives software teams a clearer target than ad hoc compatibility layers.
• Open-source chip work becomes useful when multiple companies can build from the same base.

What happens next

The next question is whether China can turn these achievements into widely used products outside research circles. The answer will depend on three things: how fast the software ecosystem matures, how many vendors ship real silicon, and whether performance gains hold up under commercial workloads.

If Xiangshan keeps attracting chip teams and Ruyi keeps expanding support for standards like RVA23, China could become one of the most important test cases for open-source silicon at scale. If that happens, the real story will not be a single announcement from Beijing. It will be the moment RISC-V stops being the alternative architecture people talk about and starts being the architecture they actually ship.

For now, the takeaway is straightforward: watch which companies announce products built on Xiangshan over the next 12 months. That will tell us whether this is a research milestone or the start of a much bigger commercial shift.