What is RISC-V?

Recently, RISC-V has gained significant attention, with many industries in China focusing on and joining this movement. So, what exactly is RISC-V? Let’s take a look.What is RISC-V?

1.Background and Technical Philosophy

  1. Origins and Design Goals

RISC-V was developed in 2010 by a team led by Krste Asanović and David Patterson at the University of California, Berkeley, aiming to replace the outdated MIPS and SPARC instruction sets used in educational settings.

.

  • Core design principles: Reduced Instruction Set Computing (RISC), utilizing fixed-length instructions and a modular architecture (with only 47 base instructions), significantly lowering the hardware development threshold.

  • Key Innovations: Completely open-source and royalty-free, allowing for the free expansion of the instruction set, breaking the patent barriers of x86/ARM.

  • Commercialization Milestones

    • In 2015, SiFive was established as the first RISC-V company, promoting the transition of the instruction set from academia to industry.

    • In 2018, policy support in China accelerated implementation: the RISC-V Industry Alliance was established in Shanghai, Alibaba acquired Zhongtianwei to form “Pingtouge,” and released the Xuantie C910 (the first mass-produced high-performance RISC-V processor).

    📈 2.Development History and Commercialization Process

    Phase 1: Breakthrough in the Embedded Market (2018-2022)

    • Application Scenarios: Focused on embedded fields such as IoT devices, industrial controllers, and wearable devices, rapidly penetrating the market due to low cost, low power consumption, and high customization.

    • Representative Products:

      • In 2019, GigaDevice launched the world’s first RISC-V general-purpose MCU, the GD32V series.

      • Espressif Systems fully transitioned to self-developed RISC-V architecture, launching eight chips including the ESP32-C3, saving on ARM royalty expenses.

    • Shipping Milestones: In 2022, global chip shipments exceeded 10 billion units, taking only 12 years (compared to 21 years for ARM and 30 years for x86).

    Phase 2: Expansion to High Performance and Multiple Fields (2023-Present)

    1. Technical Performance Breakthroughs

    • Server-Class Processors: Alibaba’s Xuantie C930 has a clock speed of 3.4GHz, with a SPECint2006 score of 15.2/GHz, integrating an AI acceleration engine (8 TOPS computing power).

    • Automotive Chips: Yiswei’s R500A core has passed ASIL-B automotive certification, and Chipone Technology has launched ASIL-D level automotive IP cores.

    • AI Inference Optimization: RISC-V vector extensions (RVV) support efficient parallel computing, with Zhihe Computing launching the A210 chip for unified inference, reducing costs for large model inference.

  • Ecological Expansion and Standardization

    • RVA23 Specification: A unified 64-bit application processor standard to address fragmentation issues.

    • Software Ecosystem: NVIDIA has ported CUDA to RISC-V, with Linux/Android gradually adapting, but the automotive-grade AUTOSAR toolchain still needs improvement.

    🌐 3.Current Status and Market Landscape (2025)

    1. Market Size and Penetration Rate

    • Global Shipments: Expected to exceed 10 billion units in 2024, with China accounting for over 50%.

    • 2031 Forecast: Shipments are projected to reach 20 billion units, with significant shares in consumer electronics (39%), automotive (31%), and data centers (28%).

    2. Corporate Ecosystem and Commercialization Progress

    • Leading Public Companies:

      • Company Direction Representative Products/Results
      GigaDevice Dual Architecture Parallel (ARM+RISC-V) GD32V Series MCU
      Espressif Systems Fully Self-Developed RISC-V IP ESP32-P4 (Dual-Core AI Chip)
      Yiswei Computing Automotive/High-Performance Computing R500A Core (ASIL-B Certified)

    • IPO Wave: Yiswei is aiming for an IPO on the Hong Kong Stock Exchange, while Qinheng Micro (with an IPO fundraising of 932 million) has been profitable for three consecutive years.

    • Startups Focusing on AI: Zhihe Computing and Jindie Shikong are developing RISC-V architecture AI server chips.

    3. Cutting-Edge Technology Applications

    • AI Inference: RISC-V optimizes operator efficiency through scalability, with ZTE Microelectronics launching RISC-V+GPU heterogeneous AI integrated machines.

    • Automotive Electronics:

      • Weak Ecological Scenarios: Body control (Great Wall’s Zijing M100 MCU), charging modules.

      • Strong Ecological Challenges: Intelligent driving requires overcoming operating system adaptation (Harmony/Android) and toolchain maturity.

    • High-Performance Computing:

      • The Xiangshan processor (Chinese Academy of Sciences) Kunming Lake V2 architecture achieves 18.5 points/GHz, competing with ARM A76.

      • Tenstorrent’s Ascalon chip exceeds Intel’s Sapphire Rapids in integer performance.

    ⚠️ 4.Challenges and Future Trends

    1. Current Bottlenecks

    • Ecological Fragmentation: Customization in the embedded field leads to a lack of unified software and hardware standards.

    • Insufficient Validation in High-Performance Scenarios: Server/PC chips lack large-scale implementation cases, often remaining at the “PPT benchmark” stage.

    • Automotive Certification Barriers: AEC-Q100/ISO 26262 certification cycles are long (2-3 years), with incomplete IP support.

  • Future Directions

    • AI-Driven Ecological Reconstruction: Utilizing modular architecture to adapt sparse large models (such as DeepSeek-MoE), seizing the edge inference market.

    • Accelerated Domestic Replacement: China aims to achieve chip independence through RISC-V, targeting a 30% penetration rate for domestic servers by 2025.

    • Global Competition and Cooperation: The RISC-V International Foundation promotes standard unification, with Chinese companies leading application innovation, while Europe and the US excel in foundational IP and toolchains.

    💎 Summary

    RISC-V, with its open-source freedom and modular flexibility, has evolved from an academic prototype to a billion-unit industry force within a decade. China has become the core engine of its ecological expansion, achieving a commercial closed loop in the MCU and AIoT fields, and is now tackling the “hard battlefield” of high-performance computing and automotive electronics. Whether it can truly compete with x86/ARM depends on the successful launch of benchmark terminal products (such as RISC-V PCs/servers) and global ecological collaboration efficiency, with AI and software-defined vehicles (SDV) potentially serving as key springboards for its “third of the pie”..

    To get more IoT news, please click the link below and follow us!

    Leave a Comment