Shanghai, China, November 8, 2024- AI chip R&D and basic computing platform company Aixin Yuanzhi announced that Dr. Qiu Xiaoxin, founder and chairman of Aixin Yuanzhi, was invited to attend the Global CEO Summit recently held by Aspen Core, a well-known media group in the global electronic technology field, and delivered a keynote speech, sorting out the latest trends in semiconductor development in the era of artificial intelligence, as well as the challenges and opportunities of AI chips under the background of accelerated integration of cloud edge end.

At the 2024 Global Electronic Achievement Awards held at the same time, Aixin Yuanzhi was awarded the "Most Promising Artificial Intelligence Technology Enterprise of the Year" for its technological innovation and excellent product application performance in the field of AI, and Aixin Tongyuan AI processor was awarded the "Innovative Product Award of the Year".

Big models promote the interconnection of all things, AI capabilities sink to the edge

With the gradual popularization of large-scale models, artificial intelligence is bringing disruptive changes to people's production and life. Qiu Xiaoxin said that in order to deeply benefit mankind, AI needs to further sink to the edge and end. When the data center accelerates, the edge side intelligence has more advantages in privacy protection, low latency response and cost control of equipment, which is also the foundation of the Internet of Everything. "The emergence of mobile Internet before has greatly promoted the development of semiconductor industry and technology, especially the development of mobile devices that adapt to high efficiency and low energy consumption. The explosion of AI provides new opportunities for the semiconductor industry and technology, that is, under the condition of controllable performance, cost and power consumption, let AI really run on edge side and end side devices."

In the past three years, artificial intelligence has entered a stage of development where a hundred flowers bloom and a hundred models compete. Artificial intelligence technology has begun to surpass humans and can largely replace people in doing some logical reasoning work. Qiu Xiaoxin believes that various phenomena indicate that the real landing of AI infrastructure is imminent. From the perspective of the semiconductor industry, how to truly achieve cloud edge intelligent integration of AI chips is a common issue that related industries will face for a long time in the future, and Aixin Yuanzhi's preliminary answer is edge side artificial intelligence chips.

Aixin Yuanzhi believes that edge AI chips are an important pivot to leverage the intelligent world. Qiu Xiaoxin pointed out that, compared with the cloud, the edge side AI chip also pays attention to the processing of perception signals in addition to computing. "One of the most important functions of edge computing or the end side chip is to digitally transform the information in the physical world in the way of artificial intelligence." At the same time, the edge side reasoning chip can improve the reasoning efficiency on the basis of controlling costs and energy consumption, and is also more conducive to training and reasoning data in the vertical field.

In the era of edge intelligence, more economical, efficient, and environmentally friendly AI chips will receive more attention from enterprises, "Qiu Xiaoxin emphasized. In other words, high computing power, low cost, large bandwidth, and low power consumption are the true directions for the actual research and development design of the AI chip industry.

Reduce costs, increase efficiency, and expand scenarios. In the era of intelligence, AI processors are needed

As chips become larger and training and inference costs increase, companies need chip products that are more in line with practical applications and AI processors that are adapted to the era of artificial intelligence. Therefore, Aixin Yuanzhi officially launched the Aixin Tongyuan AI processor to the public this year. Qiu Xiaoxin stated in his speech that the core highlight of Aixin Tongyuan AI processor lies in its innovative operator instruction set and data flow microarchitecture design concept. The processor adopts a highly programmable data stream microarchitecture at the bottom, which cleverly balances the relationship between energy efficiency and computing power density, enabling the processor to demonstrate excellent performance and efficiency when executing complex AI tasks. Meanwhile, its flexible architecture design ensures the completeness of the operator instruction set, providing solid support for various AI application scenarios and greatly expanding the application boundaries of AI technology.

It is worth mentioning that due to the early prediction of the development of Transformer network architecture, Aixin Tongyuan AI processor natively supports Transformer, which also ensures that its energy efficiency is ahead of higher end AI chips and more suitable for the landing application of edge large models.

Based on the two core technologies of Aixin Zhimou AI-ISP and Aixin Tongyuan NPU, Aixin Yuanzhi is continuously landing in the fields of smart cities, intelligent driving, and edge intelligence. At the meeting, Qiu Xiaoxin used the landing application of Aixin Yuanzhi in the field of intelligent driving as an example to demonstrate the actual effect of Aixin chip products applied to electronic exterior rearview mirrors (CMS) on site. Based on AI's subversion of traditional image processing, Aixin Zhimou AI-ISP truly achieved the visual effect of "night as white as day". At the same time, Aixin Tongyuan NPU improved the performance of image processing. With the support of this entire set of technology, intelligent driving accessories such as CMS can obtain better visibility in extreme scenarios, thereby improving the safety of intelligent driving.

At the end of the speech, Qiu Xiaoxin stated that the development of technology constantly poses challenges to industry problems, such as how to reduce the dependence of AI inference performance on DDR bandwidth, whether Chiplets have a future on the edge side, and how to break low-end competition through technological innovation... "However, survival and innovation should not be a multiple-choice question. Artificial intelligence will definitely move towards the era of ubiquitous intelligence, and sustainable and benevolent development can truly promote the universality of AI.

Regarding Aixin Yuanzhi:

Founded in May 2019, Aixin Yuanzhi Semiconductor Co., Ltd. is committed to building a world leading AI perception and edge computing chip, serving the huge edge and end device markets such as smart cities, intelligent driving, edge computing and robots. The company has independently developed two core technologies - Aixin Smart Eye AI-ISP and Aixin Tongyuan Hybrid Precision NPU. In the past five years, it has completed the research and mass production of multiple generations of intelligent chip products. Aixin Yuanzhi adheres to the mission of "inclusive AI, creating a better life", with the vision of "building a world-class perception and computing platform", committed to becoming the digital gateway to the physical world, providing a basic computing power platform for artificial intelligence for different industries, and making intelligence truly accessible.

This is reported by Top Components, a leading supplier of electronic components in the semiconductor industry

They are committed to providing customers around the world with the most necessary, outdated, licensed, and hard-to-find parts.

Media Relations

Name: John Chen

Email: salesdept@topcomponents.ru