GuShiio.com Wen| New Position Pro
The relevant event that occurred at a time that most closely matched the starting point of Ali’s rising cycle was that DeepSeek began to explode around the world. The recent lows of Old BAT were on January 13 (the U.S. AI chip restriction order), and then they collectively rose for some time (DeepSeek exploded against the chip-only theory). In the process, Nvidia plunged (which is gradually recovering). DeepSeek brought new ideas are reshaping the valuation logic of the global AI industry in the short term.
In BAT, Ali has a relatively stable e-commerce business in its old bank. Taking advantage of this wave of attention, the market’s confidence in Ali during this period may be mainly aimed at the AI business that Ali has already made achievements, such as To B’s Alibaba Cloud, To C’s Quark, Upstream Pingtou Brother, and model-level technical skills.However, none of these really conveyed to the market the new imagination that belonged to Ali alone. Before DeepSeek, Ali had already done so.
If you want new imagination, you need to start with Ali’s weaknesses in AI.In fact, in 2024, taking the development of Tongyi app as a typical example, not all Alibaba’s AI to C business will go smoothly. In the past period of time, Ali’s new layout in its own AI to C business has accurately targeted the past blind spots and made just adjustments to the focus, structure, and personnel of the business.
Just as these foreshadows were laid one by one, but before they were detected by the market, Apple quietly sent another wave of east wind. Just before the opening of the market on the 11th, The Information reported that Apple Ali had reached a cooperation to develop artificial intelligence functions for China iPhone users. After the news broke, Ali’s overall share price continued to rise that day, eventually closing at US$112.78.
In the past year, the high-profile issue of who can win Apple’s AI orders in China has remained unresolved. For the domestic AI industry, this is a key node. For Ali, it symbolizes that the AI to C business may officially become its next imaginative curve.
After all, in order to continue to shape the market’s confidence in its AI business, Ali cannot just have the original things.
01. Just sorted out the AI to C architecture
In mid-January, Ali’s main AI to C (including agents) businesses, which had previously developed in uneven conditions, were integrated and handed over to Wu Jia.
[Intelligent Information]’s business includes Quark, Tongyi To C team (just transferred from Alibaba Cloud to Intelligent Information at the end of the year), UC browser, etc. Quark’s results last year are obvious to all. According to Qimai data, its bean buns, which exceed bytes, have become the number one AI application in downloads in 2024; Tongyi app, as previously analyzed by “New Position”, has been overtaken in the past year and missed the top three in 2024 chatbot.
On the other side, Tmall Elf is currently Alibaba’s more representative hardware Agents business and is the core business of the [Intelligent Internet] business group. On January 16, according to the news of “Intelligent Emergence”, Ali handed over the Intelligent Internet business group to Wu Jia, president of the Intelligent Information Business Group, and Tmall Elf’s hardware team is already working integrated with the Quark product team.
There are also some core AI to C products scattered across other architectures. For example, Independent Nail ranked fourth in AI application downloads last year, and the top four AI products were Ali-Byte-Byte-Ali. Alibaba Byte was also inseparable in AI to C applications last year.
The reason why these businesses have developed uneven at present is that not all businesses occupied good market positions before the era of big models.
For example, Ali’s mature AI to C business has nails and quarks. These two products have accumulated quite rich business ecological richness and product power long before the big model era, and are already the leading products in their respective fields., should have a higher right to speak independently. Perhaps because of this, the AI model behind Quark was previously developed by Quark, and Dinghao even got rid of the cloud nail and became independent in 2023, focusing on doing his own business well.
However, Tongyi C-terminal is a product completely born in this era. It was previously attached to Alibaba Cloud and later came to the organizational structure alongside Quark. It is a very important business that still needs to be improved; as for the entire home smart interconnection industry where Tmall Elf is located, as mentioned in previous articles in “New Position”, this field has always been dominated by home appliance giants.
Previously,”New Position” also expressed a view. When a big factory had a good development and a high market share in the era before the outbreak of the big model, then its trump model and the newly born To C product The influence will be relatively compared.Judging from Tongyi’s C-terminal process, C-terminal products may even be restricted to some extent by cloud services.
So if you need to further develop AI to C (including Agents) related businesses,Ali needs a lead that can reasonably connect AI to C businesses and cloud businesses, while maintaining the autonomy of mature and soon-to-mature businesses.
Construction started after the Spring Festival holiday, and a new round of personnel changes in the company is underway. Ali ushered in this lead.
02. Start balancing technology and To C products
On February 6, First Financial News, Alibaba hired Singapore artificial intelligence scientist Steven Hoi as vice president of AI to C. Xu Zhuhong himself also confirmed and posted the source:
According to information from the media Shen Xiaofei, Xu Zhuhong had joined Alibaba before the Spring Festival and was under Wu Jiazhi, vice president of Alibaba Group and president of the Intelligent Information Business Group.
In addition to interface news, Xu Zhuhong is responsible for the multimodal basic model of the AI To C business and basic research and application solutions related to Agents, improving the end-to-end closed-loop capabilities of Alibaba’s AI application C-end products in model-combination applications.
All in all, Xu Zhuhong’s position in Ali’s organizational structure is most likely to be the vice president of intelligent information and reports to Wu Jia, president of intelligent information; and the content he is responsible for can be summarized as the basic model/solution construction of Ali AI to C and Agents related businesses.
Combined with the personnel changes in mid-January, as well as Alibaba’s previous AI to C business organizational structure and development status, this arrangement now also means that Alibaba is currently more clear about its business organization in this area.——Ali’s main AI to C and Agents businesses have been integrated and need to link and balance the roles of technology and product sides. So the business division of AI to C and Agents ‘basic model construction is indeed a good answer.
When the team is established, this is equivalent to an intermediate role and middle platform of the internal business model, which is used to balance the technology side with the AI to C and Agents application side, mature business and immature business.
Xu Zhuhong’s previous relevant experience is very consistent with Ali’s current development vacancy needs in this area. It is reported that Xu Zhuhong served as vice president of Salesforce Group, the world’s largest enterprise software manufacturer, and founding director of the Asian Research Institute, leading a number of artificial intelligence research projects and application products.
03. Prepare to conceive Agents business
The cooperation between Apple and Ali has not received a response from both parties so far (February 12). However, public opinion and the market have confirmed the rationality of this cooperation.Many people in the industry have already realized that Ali is indeed the best answer for Apple AI in China.
From Apple’s perspective, Ali has the technical power it wants, its caution about data security, its vision for AI and cooperative ecology, and sufficient service capabilities. From Ali’s perspective, no one doesn’t want Apple’s orders to be too general,The core significance of being able to win Apple orders is to contribute an AI to C hardware cooperation paradigm and benchmark case for Alibaba and the industry.
For Ali, the most representative AI to C hardware business is the above-mentioned Tmall Elf that has just been integrated into Quark. Cooperation with the world’s top To C hardware products with AI functions will naturally feed back to Tmall Elf. This business may even open up more AI to C hardware businesses.
The AI functions required by AI to C hardware not only include generalizations, but also consider agents-related functions. Agents are generally believed by the industry and one of the most imaginative directions for the implementation of To C in the AI industry.
In the past two years, the AI to C hardware direction has been in a state of chasing hot spots but lacking benchmarks. From the very popular AI mobile phones a year ago to the current connection of major new energy and mobile phones to DeepSeek, it is not difficult to see that there is still a lot of confusion in the direction of implementation.
In fact, the biggest difficulty in implementing Agents at present is not technology. A previous article in “New Position””Service +Agent, Zhi Xiaobao Reconstructs Alipay” mentioned a point:The biggest pain point for users using AI Agents is the diversity and integrity of services within the ecosystem, as well as interconnection.
Whether it is an AI mobile phone or a large new energy access model, if they are all within the ecosystem of the same interest, then the Agent’s instructions will of course be unimpeded. However, AI helps people do things only in the same ecosystem. For the C-end users, the efficiency increase in work and life is still not a qualitative leap, but the cross-interest Agents instructions will involve the agreements of all stakeholders.
What AI to C hardware lacks now is precisely this kind of cooperation agreement paradigm that connects multiple ecosystems. Who should study this paradigm? In China, Ali and Apple are the best pair of answers. I even believe that this combination is more imaginative than the combination of OpenAI and Apple.
Apple has a hardware ecosystem that spans multiple scenarios at work and home, and will not even rule out the possibility of Apple building cars in the future. Ali has a partner ecosystem that spans entertainment, local life, office, e-commerce, etc., which is almost the entire scenario for C-end users, so that there is no difficult business in the world. At this moment when AI Agents need to be interconnected, the value is brought into play.
Of course, these are all slightly distant future words.
