Amazon Global Selling: A Decade of Growth in a Vast Market
Dec 17, 2024 05:43 PM
Din Tai Fung and the Globalization of Chinese Cuisine
Dec 03, 2024 08:26 PM
Podcast Marketing, A Useful Tool for Companies Going Overseas
Dec 02, 2024 02:15 PM
The Chinese intelligent chip industry is currently in a period of vigorous growth, with estimated market size of CNY 288.19 billion by 2027.
chip
With the rapid development of technology, the intelligent chip industry has become the core of the information industry, leading the key force in the new wave of technological revolution and industrial transformation. Chips play a vital role in handling a large number of computing tasks in various artificial intelligence applications and are often referred to as AI accelerators or compute cards. They involve the fusion of computer science and semiconductor chip fields, where the former deals with efficient intelligent algorithms, and the latter focuses on effectively implementing these algorithms on silicon wafers, ultimately creating physical products that can integrate with supporting software.
Currently, artificial intelligence chips can be roughly divided into CPU, GPU, FPGA, ASIC, neuromorphic chips, as well as cloud AI chips, edge AI chips, and terminal AI chips based on their technological architecture and their position in the network. They can also be categorized as training chips and inference chips based on their practical goals. These different types of chips, while catering to different application scenarios, collectively contribute to the flourishing development of the artificial intelligence chip industry.
The Chinese intelligent chip industry is currently in a period of vigorous growth, with the continuous advancement of third-generation semiconductor technology, and the increasing implementation of computational power and terminal applications. Although China's artificial intelligence chip development is still in the early stages of its lifecycle, the broad development space in technology research and development and terminal applications presents promising prospects. With the joint promotion of policies, markets, and technology, it is expected that by 2024, China's artificial intelligence chip market size will exceed CNY 100 billion, and by 2027, this number will further reach CNY 288.19 billion.
At this opportune moment, this article will introduce 12 leading Chinese intelligent chip companies and their performance in terms of technological innovation, market competition, and industry development, offering readers an entirely new and exciting perspective.
1. Cambricon (Chinese: 寒武纪)
Founded in 2016, Cambricon is China's most representative domestic AI chip manufacturer, mainly focusing on the research, design, and sales of AI core chips for various cloud servers, edge computing, and terminal devices. Its main product offerings include cloud-based AI chips and accelerator cards, training machines, edge AI chips and accelerator cards, and terminal intelligent processor IPs, along with corresponding software development platforms such as the SiYuan (Chinese: 思元), XuanSi (Chinese: 玄思), and other chip/accelerator card product series.
The company is a scarce provider of AI computing power chips in China, with a comprehensive product layout that offers a series of intelligent chips and platform-based basic system software capable of serving cloud, edge, and terminal applications. In the cloud product line, the company has successively introduced the SiYuan 290 and SiYuan 370 chips, along with corresponding cloud-based AI accelerator card series and training machines. As for the edge product line, the company has launched the SiYuan 220 chip and edge AI accelerator cards, which have been adopted by several leading enterprises in edge computing scenarios, achieving cumulative sales of over one million units since their release. In the area of IP authorization and software, the company has released the 1A and 1M series intelligent processors designed for terminal applications and authorized their use by customers in their products.
2. Horizon (Chinese: 地平线)
Horizon is a leading provider of high-performance intelligent driving computing solutions. As a pioneer driving the commercial application of intelligent driving in the Chinese passenger car field, Horizon is committed to developing cutting-edge hardware computing platforms and open and user-friendly software development tools through a forward-looking technological approach, providing core technical infrastructure and an open and prosperous software development ecosystem for the transformation of the intelligent automotive industry, offering users an unparalleled intelligent driving experience.
In 2020, Horizon officially started the year of mass production of China's automotive intelligent chips, achieving a breakthrough from 0 to 1. As of today, the cumulative shipment of Horizon's chips has exceeded 1.5 million pieces, and they have signed over 70 models for mass production projects with more than 20 automakers, jointly exploring the value from 1 to N.
3. Kunlunxin (Chinese: 昆仑芯)
Formerly known as Baidu's Intelligent Chip and Architecture Department, Kunlunxin completed independent financing in April 2021. The company's team has been deeply involved in AI acceleration for more than ten years and is an AI chip enterprise with rich accumulation in architecture, chip implementation, software systems, and scene applications. Adhering to the mission of "making computing more intelligent," Kunlunxin focuses on building powerful, easy-to-use, and high-performance general AI chips. It has already achieved mass production and landing applications of two generations of general AI chip series products, with large-scale deployment in the internet, smart industry, smart finance, and other fields, empowering a thousand industries with AI computing power.
4. T-Head (Chinese: 平头哥半导体)
T-Head Semiconductor Co., Ltd., established on September 19, 2018, is the semiconductor chip business entity wholly owned by Alibaba Group. T-Head has an all-stack product series that covers data center AI chips, processor IP licensing, etc., achieving full coverage of chip end-to-end design links.
In 2019, T-Head released the data center AI inference chip, Hanguang 800, based on 12nm process technology with 17 billion transistors and a peak performance of 820 TOPS. In the industry-standard ResNet-50 test, the inference performance reached 78,563 IPS, and the energy efficiency reached 500 IPS/W. Hanguang 800 adopts T-Head's self-developed architecture, achieving performance breakthroughs through collaborative design of software and hardware. The self-developed AI chip software development kit allows Hanguang 800 to deliver high throughput and low latency for deep learning applications. Hanguang 800 has been successfully applied in data centers, edge servers, and other scenarios.
5. Enflame (Chinese: 燧原科技)
Enflame focuses on providing cloud and edge computing power products in the field of artificial intelligence, aiming to create the power base for general artificial intelligence and offering innovative AI accelerator cards, cluster systems, and software-hardware solutions with independent intellectual property rights. With its high-performance, energy-efficient, innovative architecture, and efficient and user-friendly software platform, Enflame's products are widely used in various industries and scenarios, such as the internet, smart computing centers, smart cities, smart finance, scientific computing, and autonomous driving.
6. Vastai Technologies (Chinese: 瀚博半导体)
Vastai Technologies is a self-developed high-end GPU chip and solution provider founded in December 2018, registered in Shanghai, China. The company adheres to its mission and vision of "providing vast computing power for the digital and pixel worlds" and is committed to becoming a high-end GPU chip company rooted in China and serving the world. Vastai currently has two generations of GPU chip series and has derived three major product lines: AI, rendering, and video.
Based on the VUCA unified computing architecture, Vastai achieved mass production of the cloud AI and video acceleration card series based on the first-generation chips in 2022, and it has mass-produced a series of products based on the second-generation 7nm full-featured GPU chips in April 2023.
7. Iluvatar Corex (Chinese: 天数智芯)
Shanghai iluvatar corex semiconductor Co., Ltd. is a leading provider of general GPU high-end chips and supercomputing power systems in the world. Based on its independently developed GPGPU architecture, the cloud training chip BI adopts TSMC's 7nm process, integrating 24 billion transistors and utilizing 2.5D CoWoS wafer packaging technology. The chip can support mixed-precision training such as FP32, FP16, BF16, and INT8, and it supports inter-chip interconnection, achieving a single-chip computing power of up to 147T@FP16.
As a complete solution provider with cloud-edge collaboration and training and inference combination, its system architecture, instruction set, core operators, and software stack are all independently developed and can be independently developed and evolved. Iluvatar Corex has already cooperated with important domestic industries to jointly define the design from the source, leading the way in achieving large-scale commercial mass production.
8. Corerain (Chinese: 鲲云科技)
Corerain is a global leader in artificial intelligence computing power supply, focusing on creating a high-performance, low-latency, and high-cost-effective next-generation artificial intelligence computing platform with groundbreaking data flow A chip technology. For the digital needs of industries such as energy, chemicals, power, cities, and more, Corerain provides integrated intelligent solutions of computing power, algorithms, and platforms, helping over a thousand end-users complete intelligent transformation and accelerate the application of artificial intelligence technology.
Corerain has achieved technological breakthroughs in measured chip performance through self-developed data flow technology, with a chip utilization rate of over 95%, improving the measured performance by up to 11.6 times compared to similar products. This customized data flow technology does not rely on advanced wafer manufacturing processes or larger chip sizes but enhances measured performance by controlling the flow of data to optimize performance, providing users with higher computing power cost-effectively.
9. SiEngine (Chinese: 芯擎科技)
Hubei SiEngine Technology Co., Ltd. was established in 2018 in the Wuhan Economic and Technological Development Zone. It has research and sales branches in Wuhan, Beijing, Shanghai, Shenzhen, and Shenyang. The company focuses on the design, development, and sales of advanced automotive electronic chips and aims to become a world-leading provider of overall solutions for automotive electronic chips.
The "Longying One (Chinese: 龍鹰一号)" 7nm automotive-grade intelligent cockpit multimedia chip by SiEngine is produced by TSMC and integrates "one chip, multiple screens, and multiple systems" for intelligent cockpits. It combines functions such as voice recognition, gesture control, LCD instrument clusters, HUD (head-up display), DMS (driver monitoring system), and ADAS (advanced driver-assistance systems), providing drivers with a more intuitive and personalized interactive experience.
10. MediaTek (Chinese: 联发科)
MediaTek is the world's fourth-largest fabless semiconductor company and holds a leading position in markets such as mobile terminals, smart home applications, wireless connectivity technologies, and Internet of Things (IoT) products. Each year, approximately two billion terminal products with MediaTek chips are launched globally.
MediaTek has advanced 5G solutions and has released multiple Dimensity (Chinese: 天玑) series 5G mobile chips, including flagship Dimensity 1200, Dimensity 1100, and Dimensity 1000 series, as well as global market-oriented Dimensity 900, Dimensity 800, and Dimensity 700 series. The flagship Dimensity 1200 is manufactured using a 6nm process and is a highly integrated 5G SoC (System on Chip) with advanced 5G, imaging, AI, and gaming technologies. In recent years, many popular 5G smartphones, including those from OPPO, vivo, and Xiaomi, have adopted MediaTek Dimensity 5G mobile chips. As of 2020, the Dimensity series chips have been shipped to North America, Europe, the Middle East, mainland China, Southeast Asia, and Australia.
11. Hisilicon (Chinese: 海思)
Hisilicon is a global leading semiconductor and device design company. It was formerly known as the Huawei Integrated Circuit Design Center, which started integrated circuit design and research business in 1991. In 2004, it was registered as a formal company to provide Hisilicon chips for external sales and services. Hisilicon is committed to providing leading and reliable semiconductor foundation for smart terminals in various scenarios, such as smart cities, smart homes, and smart travel, serving thousands of customers and developers.
12. NavInfo (Chinese: 四维图新)
NavInfo, founded in 2002, has become a leading provider in the fields of navigation maps, dynamic traffic information, and custom vehicle networking solutions after over 20 years of innovation and accumulation. Today, NavInfo embraces the arrival of the automotive "new four modernizations" era with a comprehensive technology development strategy. They have formed full-stack solution service capabilities for intelligent vehicles, including intelligent cloud, intelligent cockpit, intelligent driving, and intelligent chip solutions. NavInfo is dedicated to creating an "intelligent automotive brain" with core businesses such as autonomous driving system solutions, cloud service platforms, high-precision maps, high-precision positioning, and automotive electronic chips. Additionally, leveraging big data capabilities, they expand into applications for smart cities and the Internet of Everything, empowering intelligent travel and contributing to a better life, making them a trusted intelligent travel technology company by customers.
Amazon Global Selling: A Decade of Growth in a Vast Market
Dec 17, 2024 05:43 PM
Din Tai Fung and the Globalization of Chinese Cuisine
Dec 03, 2024 08:26 PM
Podcast Marketing, A Useful Tool for Companies Going Overseas
Dec 02, 2024 02:15 PM