Accelerating Domestic AI Computing Power! Opportunities Across the Entire Industry Chain from Chips to Servers

The conference emphasized the strategic position of AI, highlighting the urgent need for domestic computing power alternatives. Giants like ByteDance and Alibaba are increasing their investments in computing power, with a projected year-on-year growth of 22%-28% in computing power investments by 2025 for mobile, telecom, and Unicom sectors. Domestic GPUs such as Huawei Ascend and Haiguang Information are adapting to the DeepSeek model, breaking through overseas chip blockades. Core tracks and targets include: Computing Power Chips: Haiguang Information: Leading domestic GPU manufacturer, with DCU product performance comparable to NVIDIA’s A100. Hanwha: The Siyuan series chips are used for large model training, with orders increasing by 300% year-on-year. Servers and Equipment: Tsinghua Unigroup: Core partner of Huawei Ascend, with self-produced server capacity doubling. Zhongke Shuguang: Leading in liquid cooling technology, with AI server revenue expected to exceed 40%. Computing Power Scheduling and Operations: Siqi: Intelligent scheduling system connected to over 500,000 computing power cards. Optical Modules and Connectors: Zhongke Xuchuang (300308): 35% global market share in 800G optical modules, benefiting from AI data center upgrades. Technological Breakthroughs and Ecosystem Development: Huawei Ascend AI clusters have supported training for models with over a trillion parameters, with Shanghai’s intelligent computing cloud industry aiming to exceed 200 billion by 2027. Alibaba Cloud and Tencent Cloud are accelerating domestic adaptation, while Tianyi Cloud has achieved full-stack domestic capabilities.

Related posts

  1. Huahong Hongxin Expands Its Reach with Investment in Semiconductor Companies
  2. Essential PLC Programming Algorithms: A Beginner’s Guide
  3. Comparison Study of Commonly Used Encryption Algorithms
  4. A Comprehensive Analysis: Does China Qualify for the U.S. Exemption on Reciprocal Tariffs for Semiconductor Products like Smartphones and Computers? If So, How to Declare?
  5. Fundamentals of Tape-Out in Integrated Circuit Design
  6. Summary of 35 Major Wafer Factories in the U.S. Based on Foundry Locations
  7. Understanding HTTPS Encryption and Decryption
  8. Underlying Principles of Docker
  9. C Language Example: Determine if a Number Can Be Expressed as the Sum of Two Prime Numbers
  10. Understanding and Reflection on Edge Computing
  11. Disassembly of Waste Circuit Boards
  12. Implementing a Health Bar in C Language
  13. Research on the Autonomous Driving Chip Industry: Central Computing, Large Models, and Navigation Assistance Leading a New Wave of Innovation
  14. Successful Tape-Out of Loongson 3B6000 Processor
  15. Daily Embedded News: Successful Tape-Out of Loongson 2K3000 Processor; Intel and TSMC Joint Venture Progress; Samsung Electronics Q1 Profit Decline Pressures Chip Business
  16. Performance Soars by 4230%! Domestic AI Chipmaker Cambricon Shines as the ‘NVIDIA Alternative’
  17. Comprehensive Analysis of Magnetic Sensors and Their Applications
  18. The Battle for Computing Power Dominance: How Domestic Automotive Chips Break the “Tape-Out and Supply Cut” Curse?
  19. New Regulations on ‘Wafer Fabrication Location’ Take Effect, US Chip Giants Hit Hard by Tariffs, Accelerating Domestic Substitution Wave

Leave a Comment