Preface

Not long ago, we released the first coverage report of alibaba, focusing on tracking the company's outstanding roi for shareholders.

After the release of Alibaba's financial report for the second quarter of the 25th fiscal year, the market generally focuses more on the accelerating recovery of core businesses (such as e-commerce and cloud computing) and the high-quality characteristics of Alibaba's performance growth, leading the industry competition to provide greater value for customers, and the Alibaba management fulfilling growth commitments.

We believe Alibaba's "AI-driven" strategy progress is also worth noting.

Although China's leading internet and technology listed companies have been actively deploying in the field of AI, unlike the well-known U.S. "Big Tech" companies, their AI capabilities have not been fully reflected in the financial reports. However, based on the current actual business performance, at least one thing can be said - breakthroughs and changes are happening. According to the current trend, we believe that AI valuation will eventually be reflected in financial data, it's just a matter of time.

Therefore, we proactively track Alibaba's AI strategy layout and progress. We found that one of the "Alibaba Four Little Dragons", Quark, has made unexpected progress in the past two quarters.

We believe that this change will play an important driving role in the reassessment of the timing of Alibaba's AI potential release and investment value. Furthermore, it can be inferred that Alibaba's AI strategy has taken the lead in breakthroughs among China's internet and technology listed companies.

1. Alibaba's "AI-driven" strategy and the unique position of Quark

In today's global technology field, artificial intelligence (AI) technology has become the new core of business competition. This business battle, known as the "AI Battle", is constantly affecting the nerves of global technology companies. They are using AI technology and products as weapons to engage in fierce market competition and internal innovation.

Under the vast wheel of the times, Alibaba is also actively planning its layout.

In fact, AI technology has long been seen as one of the key factors driving Alibaba's future business growth. In April 2023, Alibaba's fundamental large-scale model Tongyiming Thousand Questions was introduced. In September 2023, Alibaba identified "AI-driven" as one of the strategic focuses.

In May 2024, Alibaba interpreted its AI-driven strategic thrust in its "2024 Fiscal Year Letter to Shareholders": first is the fundamental large-scale model, second is the cloud computing business, third is application scenarios.

In terms of specific business, the first refers to Alibaba Tongyiming large-scale model; the second refers to Alibaba's cloud computing business, and the third is representative AI applications extended from Alibaba's various C-end businesses.

In 2024, it is widely recognized as the first year of AI landing in the industry. Major domestic internet companies have successively launched AI products representing their C-end layout, such as ByteDance's Douyin, Baidu's Wen Xiaoyan, etc.

On the other hand, regarding Alibaba, we believe that Quark, hidden in the financial reports of other businesses, can represent Alibaba's C-end AI products for three reasons:

1. Quark is independently operated under Alibaba, and it started early in the layout of intelligent products in the AI field. Quark was positioned in 2018 as an intelligent search engine with AI technology at its core, placing AI technology at the core of its product strategy from the beginning. After years of investment and perseverance, Quark has taken the lead in products and technology.

Quark's comprehensive product capabilities have entered the top ranks of domestic and even global AI products. The latest data from Ai Products List released by a well-known domestic third-party data service agency, New Rankings, shows that Quark ranks fourth globally and first in China. Qimai's 'Q2 2024 iOS Strength AI Product Rankings' also shows that Quark has won the industry championship with the highest rating.

Quark leads in word of mouth among young people. Seizing young users means seizing the future of technology and business. Young people have a higher acceptance of new products and a larger user base. QuestMobile data shows that the number of students aged 24 and below in China has reached 0.261 billion. It is understood that half of Quark's users are under 25 years old, and the proportion of young people on Quark's PC side even exceeds half.

We believe that Quark, which has already risen to the top tier of global C-end AI applications, is becoming the leader in Alibaba's AI strategic C-end application field, bringing more imagination space to Alibaba's AI-driven strategy.

Through the window of Quark, we can gain a deeper insight into the value of Alibaba's AI strategy.

Second, the full embodiment of AI strength: Quark's PC-side leading edge.

First, let's briefly review Quark's development trajectory and see how it has gradually risen to become the 'ace' of Alibaba's C-end AI applications.

Since its inception, Quark has defined its search business as intelligent search, delving deep into the application of AI technology in its products.

In November 2023, Quark released the Quark Mega Model, marking its further deepening in the field of AI. At that time, the Quark Mega Model successively topped the C-Eval and CMMLU authoritative evaluation lists, with many performance indicators superior to GPT-4, reflecting Quark's advantage in large model capabilities.

In July 2024, Quark upgraded the 'Super Search Box' and launched a one-stop AI service centering on AI search; in August 2024, Quark released a new PC end. Quark quickly upgraded its multi-terminal products to products with comprehensive AI capabilities.

According to public information, Quark's PC end has upgraded a series of functions such as AI search, AI writing, AI PPT, and AI document summarization, relying on the 'system-level full-scenario AI' capability to help users complete information retrieval, creation, and summarization in one stop.

Using Quark's PC end as an example to analyze Quark's AI capabilities. Since the release of the new Quark PC end, its traffic and user base have grown rapidly, while the number of AI vertical applications continues to increase with the expansion of application scenarios. The launch of new products will further attract new users and downloads, forming a positive feedback loop, which is a typical appearance of the 'flywheel effect'.

According to the latest data from PConline, Quark's PC end ranks first in the download volume of AI applications on domestic computers in 2024, becoming the preferred PC end AI product for young people. Quark PC's download volume is significantly higher than the second-placed Bean Bag.

Analyzing the reasons for the success of Quark PC end, we have the following conclusions:

The comprehensive and deep integration of Quark's ai function is one of the core driving forces for its increased traffic this year. In the productivity tool usage scenario, a single AI plugin often only meets scattered demands, while Quark seamlessly integrates these demands through the "system-level all-scenario AI" capability, enhancing user experience and stickiness, making AI an all-round assistant for user work and study. For example, the Quark PC product integrates a full range of AI applications and rich information services, upgrading the computer to an "AI computer" without requiring hardware upgrades.

The product is outstanding enough. The Quark PC side not only has superior performance but also features convenience and ease of use. Its user-friendly interface design and convenient usage process provide a unique information collection, processing, and summarizing experience, especially in academic search, AI topic search, AI writing, and AI PPT functions.

According to the theory of the internet product life cycle, Quark's extensive monetization actions on the PC side have not yet appeared. On the one hand, including the PC side, many AI products or functions in Quark's product matrix are free; on the other hand, Quark has been very restrained in commercialization, always adhering to long-termism and craftsmanship to polish products and accumulate users. Alibaba's "user-first" spirit is well reflected in Quark. Therefore, it can be determined that Quark is currently in the mid-term stage of the fastest user growth, and its future potential commercial value release will be more remarkable.

Quark's commercial potential in the future is enormous from both data and scenario perspectives. Quark is still in a high-growth period without sacrificing user experience.

Verifying from both data and scenario perspectives, Quark has huge future commercialization potential.

Quark is still in a period of rapid development without sacrificing user experience.

According to the latest statistics from the Quantum Bit Think Tank, Quark is not only leading in user scale among domestic AI products, but also ranks first in growth rate and user stickiness in the industry. Quantum Bit think tank data shows that in January-October 2024, the new download volume of Quark is close to 0.27 billion, which is significantly ahead of the second-ranked bean bag's 0.14 billion; in terms of user stickiness, Quark's three-day retention rate exceeds 40%, and the seven-day retention rate is close to 30%, both far higher than the industry's requirements for the retention rate of excellent products.

Another point is the rich scenarios, which also determine the high conversion and growth potential of users. One of Quark's core advantages is its deep accumulation in vertical application areas. Taking the "AI Q&A" feature for learning scenarios as an example, in order to further solve the problem of users not finding questions or not understanding difficult questions, Quark comprehensively upgraded the "AI Q&A" product, making the question and answer process faster and more powerful. Quark's "AI Q&A" is also the first Q&A product to complete a comprehensive AI upgrade across the entire network. Behind this is Quark's early layout and years of deep cultivation in the learning field. Quark has long insisted on self-built high-quality learning content, owning the largest learning material database in China and applying it to model training to ensure the accuracy and reliability of AI explanations.

Deeply cultivating vertical scene applications will bring Quark new growth and commercial space. According to the "AI Productivity Tools Summer Development Report" released by Aurora's Moon Fox Data, AI productivity tools are showing a trend of rapid growth on the user side, with a significant increase in growth rate, among which Quark App leads in new summer users.

When will Alibaba start to realize the strategic value of "AI-driven"?

First, our conclusion is: Benefiting from the market's attention to AI applications, as well as the advantages accumulated by Ali's AI applications such as Quark in the long term, it helps to increase the proportion of the overall valuation of AI-related businesses in Alibaba's market cap.

The main reasoning logic is concentrated in the following areas:

1. Focus of the market on the value of AI applications

We observe that the market focus has shifted from large model capabilities to real AI application scenarios, as well as real user usage and retention.

For AI applications, transitioning smoothly from the concept and idea validation stage to the actual user scenario application stage has realized the substantial transformation of innovative ideas and technological achievements to user-scale operations, opening up the imagination space for future new commercial opportunities.

We believe that AI applications like Quark have shown their future business potential through the release of user value. The reason is that Quarks have found the PMF (Product-Market Fit), which will drive Quarks to open up new incremental markets, leading to an increase in overall user scale and usage to a new level.

(2) Discussion on Valuation Methods for AI Applications

For the valuation methods of unlisted companies, investment banks generally use market approach, income approach, cost approach, and comprehensive approach.

We believe that at the current stage of development, using the market approach for valuing AI applications is relatively appropriate, with a certain degree of rationality and fairness.

Among them, the market approach includes market multiple method, recent financing price method, and industry indicators method. The key to the market approach is to find comparable assets or benchmark companies, which have publicly available price information. Below, we will take the market approach as an example and attempt to forecast the valuation of AI applications represented by Quark.

Using the comparative valuation method: The valuation can partly refer to the latest valuation of AI applications such as Perplexity. It is understood that after the latest financing, Perplexity's valuation reached 9 billion US dollars. By proportionally calculating its valuation and monthly active users with Quark, it is predicted that the latter's valuation will be significantly higher than Perplexity's.

Using the segmented valuation method: The market categorizes Quark's functionalities into large models, search tools, and 'AI + office' applications, thus enabling the valuation of certain business segments similar to Kunlun Tech and Beijing Kingsoft Office Software, Inc. Since Quark also has products like Quark Cloud Drive with a large user base, the valuation corresponding to these should be higher than a simple sum of business segment valuations.

(3) Impact on Alibaba's Market Cap

The increase in valuation of AI applications will drive the overall valuation level of Alibaba's entire AI-related business. We determine that once the market valuation of Alibaba's AI applications accounts for a certain percentage of Alibaba's total market cap, it will trigger a reassessment of Alibaba's 'AI-driven' global strategy.

Driven by the growth momentum of AI applications, the market will continue to rethink the necessity and weight of Alibaba's 'AI-driven' global strategy. In this process, quantitative change will eventually lead to qualitative change, and the continued rapid growth of Alibaba's AI product user base will undoubtedly bring about a significant transformation in Alibaba's positioning and value in the capital markets. We look forward to the arrival of this day.