China: Chip industry relies more on license-free RISC-V instead of Arm


China wants to turn more to the free RISC-V as an architecture for chips of all kinds and thus free itself from SoftBank's arm and the associated license costs. The cost itself is not the decisive factor, rather it is the sword of Damocles that China could be denied this architecture again at some point.

China fears further restrictions

The latest escalation of US trade restrictions has further alarmed China and, above all, the companies affected. Because it is probably not the end of the measures, the spiral could quickly turn even further. At some point, restrictions could no longer primarily affect pure hardware, but increasingly also software, as the example of Google shows, which is why those affected want to prepare for the worst, if possible.

Recently, this has also included the fact that China is looking for an alternative to Arm. Hardly anything in industry today works without an arm; many chips are based on one of the many architecture levels, be it a very old variant or the latest solution. With ARM, manufacturers are free to develop something completely independently based on the architecture (Architecture License Agreement) or to use ready-made construction kits (Technology License Agreement). Ultimately, Arm collects license costs for both variants.

With the industry giants Alibaba and Tencent, China now wants to create solutions with an alternative architecture: RISC-V. Started as a small university project over ten years ago and then grew into a curiosity, in recent years RISC-V has developed into a serious alternative to ARM in some areas. The advantages: It is free, open and can be expanded quickly by anyone. There was already an informative article from the ComputerBase community:

  • Reader's article “RISC-V: Why the world needs a free Instruction Set Architecture!”

The stated advantages were already a thorn in Arm's side in the middle of the last decade, whose business model at that time was largely based on license income from their closed system and which has therefore changed slightly in recent years for fear of emerging competition. Suddenly there were industry giants like Western Digital and their (partial) exit from ARM cores to RISC-V on a scale of initially one and later up to two billion possible CPU cores per year. In the end, Arm didn't even stop at a Fud campaign against RISC-V, but after a lot of response, the company quickly withdrew it. >Arm is still overpowering, RISC-V has a chance

Arm still has the upper hand, thanks to a huge ecosystem and millions of developers with the right software, top dogs in particular are little or not move at all. Qualcomm is cooking the topic on a very low flame, Intel wants to open its foundry for RISC-V and believes that in three to five years RISC-V could be found in many embedded, IoT and automotive designs, but also in mobile and data center solutions. However, the pressure to act now felt in China is accelerating things, so the forecasts for rapid growth could well be correct.

Growth potential of RISC-V (Image: Financial Times)