ASIC is a “custom chip”—unlike GPUs that can handle various tasks, it is optimized for specific functions (such as AI inference and firewall protection), achieving 40% higher efficiency and 3-5 times lower power consumption compared to general-purpose chips.
For example, when you chat with AI, the underlying “inference computation” relies on ASICs; 5G base stations processing massive data and real-time responses in autonomous driving also depend on it.
More critically, after the U.S. imposed an embargo on high-end GPUs, ASICs have become a “shortcut” for domestic computing power breakthroughs—by 2025, China’s AI chip market is expected to reach $92 billion, with ASICs accounting for over 20%. This market share is fiercely contested by domestic companies!
1. AI Inference “Leading Group”
1. Cambricon
The leading domestic AI chip company, its 7nm chip performance matches 80% of NVIDIA’s A100. Notably, it holds a 60% market share in the inference chip market priced below 2000 yuan, with an 18% penetration rate among major internet companies.
2. Rockchip
The “designated chip supplier” for ByteDance, its RK3588 chip can perform edge computing, and next year it will launch a 20TOPS ASIC supporting multimodal models. Impressively, its smart home chip shipments have increased by 80% over two years, with a market value exceeding 70 billion.
2. Design Services “Elite Group”
1. Chipone
A “hidden champion” in domestic ASIC design, with clients including ByteDance, Alibaba, and Baidu. Currently, it holds 3.025 billion yuan in orders (a historical high), with 90% being “one-stop chip customization” services. Remarkably, its AI core technology (NPU IP) has been used in 142 chip models, with global shipments exceeding 100 million units.
2. CanSemi
Linked with SMIC, a “pioneer in domestic processes,” it has pre-received 300 million yuan in the first quarter (870 million more than last year) and is developing automotive-grade MCU chips and MRAM control chips, while collaborating with packaging factories on advanced 2.5D packaging to adapt to future Chiplet trends.
3. Automotive + Communication “Potential Group”
1. Unigroup Guowei
The “domestic leader” in automotive-grade chips, it has launched the first domestic automotive MCU to pass the highest safety certification (ASIL D), which is already being tested by several OEMs. In the second quarter of this year, its net profit reached 573 million (a 381% increase quarter-on-quarter), and FPGA chip shipments remain high.
2. Aowei Technology
A “dark horse” in communication ASICs, its 4G chip shipments increased by 50% year-on-year, and its 6nm 5G chip is set to tape out in the second half of the year. Notably, its chips have entered the AI glasses and smart wearables market, adding a new track for edge computing scenarios.
3. Lanqi Technology
A “top three” global memory interface chip manufacturer, 35% of servers worldwide use its technology (even Intel is purchasing its licenses). With a surge in demand for AI servers this year, its DDR5 chips have already been sampled, indicating promising future growth.
4. Niche Scenarios “Specialized Group”
1. Hillstone Networks
The only player in network security ASICs, this year it has equipped 70 security hardware products with custom chips to combat massive data attacks in the 5G era. It has also launched an AI security integrated machine, binding with the DeepSeek large model, with performance expected to rebound in the second half of the year.
2. Nasda
A benchmark for domestic replacement in printer chips, its independently developed five main control chips are now in mass production for Bontu printers, breaking the foreign monopoly. It is also venturing into humanoid robot chips, achieving domestic replacements for joint control and power management chips.
3. Orbbec
A “hidden champion” in 3D vision ASICs, its chips are used in consumer electronics and security monitoring, with leading 3D camera technology in China.
5. Market Trends
Large Scale: The global ASIC market is expected to reach $25.6 billion this year, with China accounting for $185 billion (approximately $26 billion), where AI-related demand is growing at 30% annually;
Accelerated Replacement: After the U.S. embargo, the proportion of domestic AI chip procurement is expected to rise from 17% in 2023 to 55% by 2027, with the next four years being a golden window period;
Technological Breakthroughs: Domestic companies have already achieved breakthroughs in 7nm processes (Cambricon) and automotive ASIL D certification (Unigroup Guowei), with advanced packaging and Chiplet technology also catching up to international standards.
Feel free to like and leave comments at the end of the article for discussion.
Special Statement: The content of this account is compiled from publicly available information for research and discussion purposes only and does not constitute any investment advice. Please do not use this as a basis for investment decisions. If there are any inaccuracies, please feel free to provide corrections. Investment involves risks, and caution is advised.
#Knowledge Sharing