Ji Zhaohui, Vice President of Marketing at AMD Greater China: The ubiquitous AI engine is the key to unlocking the potential of AI | 2025 AI Partner Conference
2025 is the first year of the explosion of AI applications. A profound technological revolution is quietly rewriting the industrial landscape. At this critical juncture, the industry faces core questions: How can we bridge the gap between AI technology and large-scale applications? Where will the next disruptive AI super app be born?
On April 18, the 2025 AI Partner Conference hosted by 36Kr was grandly inaugurated at the Modu Space in Shanghai. The theme of this conference is "The Arrival of Super Apps", focusing on the disruptive transformation of all industries by AI applications. The conference is divided into two major chapters: "The Arrival of Super Apps" and "Who Will Be the Next Super App", covering seven major topics such as "Growing Up in the AI World" and "In 2025, Focus on Super Apps in AI". It includes more than 10 keynote speeches, 3 round-table discussions, and the release of two lists of excellent AI case enterprises, deeply analyzing how AI technology reconstructs business logic and reshapes the industrial landscape, and exploring the infinite possibilities brought by AI super apps.
On that day, Ji Zhaohui, the vice president of marketing for AMD Greater China, gave a keynote speech titled "Unleashing AI Potential and Building Super Apps Together".
The following is the content of Ji Zhaohui's speech, edited by 36Kr:
Respected Teacher Da, good morning, dear AI Partners! It really takes a lot of courage to give a speech after Teacher Da and Professor Liu. The threshold for today's sharing is indeed quite high. However, I will change the perspective and introduce to you the work AMD has done in the AI field from the hardware level.
First, let me briefly introduce AMD through a documentary. At AMD, we firmly believe that high-performance and adaptive computing are the cornerstones of human development. Based on this concept, AMD, leveraging its advantages in high-performance and adaptive computing, continuously tries to break through technological limits and is committed to solving many major challenges faced by humanity. For example, in Europe, we assist in interstellar exploration projects; in the Pacific region, we use AI technology to monitor the process of global warming and help maintain the healthy survival of fish; in the Asia-Pacific region, we optimize the transportation system through AI technology to improve traffic efficiency.
AMD's products are widely used in many fields such as cloud computing, healthcare, industry, communications, automotive, personal computers, and gaming. Perhaps you are more familiar with AMD's products in the gaming and PC fields, but today I will focus on the AI field. In the cloud computing scenario, when you send WeChat red envelopes, there may be an AMD processor working behind the scenes; in the medical field, minimally invasive surgical robots require extremely high precision and reliability of equipment, and AMD's embedded processors provide solid support for this; in the automotive industry, the in-vehicle systems of brands such as Tesla and Geely are also equipped with AMD's technology.
Dr. Lisa Su, the CEO and chairman of the board of AMD, once said that AI is the most important technological revolution in the past 50 years, and its influence even surpasses the birth of the Internet. In the next decade or several decades, human innovation and development will largely revolve around AI. In 2024, the awarding of the Nobel Prize in Physics and Chemistry to scientists in the AI field, although controversial to some extent, undoubtedly reflects the development trend of AI technology. AI will reshape various industries as well as our work and lifestyle, and its far-reaching impact may far exceed our current imagination.
Today, my speech topic is "Super APP". We believe that current super apps must be empowered by AI. To meet the AI computing power requirements in different scenarios, better serve various applications, and fully unleash the potential of AI, we need to build ubiquitous AI engines. As a company focusing on hardware and processor research and development, AMD's business covers data centers, embedded devices, and terminal devices, achieving full-stack AI empowerment from the cloud to the edge to the end. In the semiconductor field, few companies can achieve this.
Looking back at the era of mobile Internet, super apps like WeChat and Alipay attracted a large number of users with their one-stop, multi-functional features, forming a powerful super closed-loop ecosystem. We believe that the entry points for AI super apps are likely to appear on AIPC (Artificial Intelligence Personal Computer) and AI mobile phones. According to a research report released by IDC at the end of last year, in the next few years, human interaction with AI will become increasingly frequent and even indispensable. In the short term, AI mobile phones and AIPCs will become the main interfaces for us to interact with AI in our daily lives. Among them, AI mobile phones will mainly serve as the AI entry points for consuming life and entertainment content, while AIPCs will play a key role in productivity, work, and learning content creation. It integrates various functional requirements such as office work, creation, entertainment, and social interaction. Especially in the fields of office work and creation, the depth of its productivity scenarios is irreplaceable. For example, although mobile phones can perform simple photo editing, they are difficult to meet the requirements of complex video editing, which is exactly the area where AIPCs excel. In addition, the real-time response of super apps depends on the real-time inference of edge-side AI and edge-side models. When AIPCs perform edge-side data processing and model inference, they can also take into account the privacy protection issue that users highly value.
Next, I will introduce to you through three cases how AMD, as a hardware manufacturer, is preparing for the arrival of AI Super Apps by collaborating with ecosystem partners.
On the Windows platform, AMD has carried out in-depth cooperation with Microsoft. Microsoft released a new AIPC standard last year - Windows 11 AI+ PC. This standard specifically proposes that to welcome AI super apps, traditional PCs need to have a dedicated NPU (Neural Processing Unit) optimized for AI applications built in and have at least 40 TOPS of AI computing power. For this reason, AMD launched the third-generation AIPC processor - AMD Ryzen AI processor. This is the first processor to directly name "AI" in its product name. It not only has a powerful 12-core, 24-thread CPU and an integrated graphics card but also is equipped with a neural engine with 50 TOPS of AI computing power. At the same time, Microsoft launched new AI features in the Windows 11 operating system and announced at the press conference that it has integrated more than 40 small edge-side models into the Windows system. Currently, the Windows 11 system with AI functions has begun to be gradually pushed to the public. For example, the RECALL function can help users recall the operations they have performed on the computer. Users can find the required content through simple voice or inputting prompt words with the help of AI image search and text search; the CLICK TO DO, IMAGE SUPER RESOLUTION, and the newly launched LIVE CAPTION real-time subtitle function can perform real-time translation during live broadcasts or when watching foreign-language movies and TV shows without subtitles, greatly improving the user experience. From the operating system to the hardware, we are continuously improving for the arrival of AIPCs and super apps.
At the operating system level, in addition to the above function updates, to help ISV (Independent Software Vendor) partners develop AI applications on the new AIPC platform, Microsoft launched WCR (Windows Copilot Runtime), which provides a rich set of API modules for developers to directly call during the application development process, such as realizing functions like object erasure, super resolution, and OCR. Developers can easily call more than 40 small edge-side models built into the Windows system.
Moreover, the development of super AI apps cannot be separated from large models, especially the deployment of edge-side models. On March 18, AMD held an AIPC Innovation Application Summit in Beijing, inviting outstanding Chinese ISV partners to showcase their achievements. Teacher Da once mentioned that DeepSeek attracted much attention during the Spring Festival. Nowadays, developing AI applications often cannot do without DeepSeek. Among AMD's AIPC products, the Strix Halo processor is a "monster-level" product. As an APU, it not only has an advanced CPU and an integrated graphics card with performance equivalent to that of a discrete 4060 or 4070 graphics card but also is equipped with an NPU with 50 TOPS of computing power. Its unique UMA architecture supports 120G of memory, which can be freely allocated to the graphics card or the CPU. By allocating 96G of memory to the graphics card, this laptop can have the performance of a 4060 or 4070 graphics card with 96G of video memory, bringing infinite possibilities to application development. At the summit, ISV partners successfully deployed the 236-billion-parameter DeepSeek model on a laptop equipped with the Strix Halo processor, achieving an average inference speed of 12.25 tokens per second. Chinese ISV enterprises are very powerful, constantly pursuing technological iterations, and are committed to deploying larger-scale models on the edge side to improve performance and smoothness. We look forward to more ISV partners joining hands with AMD to explore more innovative possibilities on edge-side AIPCs.
Since March last year, AMD has established the AIPC Application Innovation Alliance in China. Currently, more than 100 outstanding ISV partners have joined, jointly promoting the development of Chinese AIPC software and hardware. If any partners on-site or in the live broadcast room are interested in participating, you are welcome to join us. The alliance has carried out a variety of cooperation activities, including market brand cooperation and promotion. We also plan to discuss joint brand marketing with Teacher Da. At the same time, we will bring software partners and OEM hardware manufacturers together and organize regular cooperation negotiations to promote cooperation between the two parties. In addition, we will also help outstanding ISV partners expand their applications overseas and explore overseas markets such as Japan.
Just now, the video showed the achievements of three ISV partners and AMD in jointly promoting Chinese AI innovation. We believe that Agentic AI will be an important implementation path for AI-based super apps. The video just now showed the demonstrations of the early AI Agent products of three ISV partners. Here, we call on more partners to join us in promoting the construction of the Chinese AI ecosystem, accelerating the development of AIPC software and hardware, welcoming the moment of innovation, and jointly promoting the early arrival of the super app era.
That's all for my sharing today. Thank you!