Alibaba, Once Focused on Business-Oriented AI, Has Undergone a Transformation!

As the year nears its end, Alibaba has made an intensive push in deploying AI products for consumers (C-end) over two consecutive days, launching two major offerings: the public beta version of the QianWen App, a 'personal AI assistant capable of chatting and handling tasks' that fully competes with ChatGPT, and the LingGuang App, an AI assistant based on full-code generation that specializes in creating natural language-generated mini-applications within 30 seconds.

The QianWen App surpassed 10 million downloads within a week of its launch, overtaking both ChatGPT and DeepSeek to become the fastest-growing AI application in history. According to official data, the LingGuang App has seen users create 3.3 million 'flash applications' within two weeks of its release.

In a short span, Alibaba has secured pivotal positions in two critical areas: personal life entry points and individual productivity tools, marking a notable phenomenon in the AI product landscape at the end of 2025.

Examining the global AI-to-Consumer (AI To C) market, competition has reached a fever pitch. ChatGPT, leveraging its first-mover advantage and mature technological experience, has deeply entrenched itself in user consciousness. In the Chinese market, it's akin to a 'Warring States era' with numerous competitors vying for dominance. ByteDance's Doubao has shown impressive performance, surpassing DeepSeek to top the monthly active user rankings. Products from major players like Tencent's Yuanbao, as well as offerings from innovative startups like Kimi, have undergone multiple iterations, each amassing a sizable and highly engaged core user base.

Historically, Alibaba's AI strategy was predominantly focused on the business-to-business (B-end) market. Now, it's clearly evolving towards a balanced approach that caters to both B and C ends. A central question emerges: Can Alibaba's 'late surge' break through amid fierce competition from industry giants and a crowded market? Another pressing concern is: Have domestic AI large-scale models truly achieved a meaningful catch-up in the global arena?

Dual-Track Strategy: Alibaba AI's Differentiated Offensive and Defensive Moves

For Alibaba, the QianWen App, evolved from the Tongyi App, is positioned as a direct rival to ChatGPT, with ambitions of becoming a personal AI life portal.

From a product standpoint, the QianWen App adheres to the mainstream paradigm of current AI assistants, featuring conversational interaction as its core and a simple, intuitive interface layout—typical traits of general-purpose AI assistants. A single command can enable the QianWen App to generate a research report or PPT in seconds.

This product, built on Alibaba's open-source large model Qwen, is positioned as a personal AI assistant capable of chatting and handling tasks, quickly igniting market enthusiasm. With over 10 million downloads in its first week of public beta, it surpassed both ChatGPT and DeepSeek to become the fastest-growing AI application in history, also ranking among the top three free apps on the Apple App Store and becoming a tech sensation.

The explosive growth of the QianWen App is no coincidence. When QianWen was primarily B-end-focused, its Qwen3-Max model excelled in internationally recognized programming and mathematics tests, even outperforming some top international models.

Previously, the CEO of Airbnb publicly stated that the company 'extensively utilizes Qwen,' and NVIDIA's CEO also mentioned that Qwen has captured a significant share of the global open-source model market.

After shifting its focus to the C-end, Alibaba plans to comprehensively integrate various life scenarios, such as maps, food delivery, ticket booking, and office services, into the QianWen App, achieving a closed loop from 'conversation' to 'task completion.'

The core of the QianWen App lies in 'integration,' while the LingGuang App targets super-individual productivity tools, with its core being 'creation.' Its most disruptive feature is the mobilization and democratization of code generation capabilities, previously exclusive to developers, in the form of flash applications. Users can generate a lightweight, interactive application, such as a travel planner or fitness tracking tool, within 30 seconds by describing their needs, supporting the output of full-modal information including 3D, audio and video, charts, animations, and maps.

Four days after its launch, the LingGuang App surpassed 1 million downloads, soaring to sixth place on the App Store's China free app chart and becoming the first AI product to exceed the download speed of ChatGPT, Sora2, and DeepSeek, demonstrating strong market demand for 'practical' AI tools.

The dual-track strategy of QianWen and LingGuang forms a comprehensive system for Alibaba's AI To C offensive and defensive moves. QianWen captures C-end traffic entry points through ecological integration, building user stickiness with scenario resources from super apps like Taobao and Alipay. LingGuang enters productivity scenarios with technological innovation, opening up a new track with its differentiated advantages of zero-code and full-modal capabilities.

This dual-track layout of life entry points + productivity tools not only avoids homogeneous competition with rivals like Doubao and Tencent's Yuanbao in general conversational scenarios but also leverages Alibaba's comprehensive strengths in technological R&D, ecological resources, and computing power support, carving out a new path in the fiercely competitive AI To C arena.

Alibaba's AI Strategy Shift: From B-End Dominance to Balanced B and C Ends

Behind the dual-track strategy lies Alibaba's strategic judgment on the development stage of the AI industry. In February of this year, Alibaba announced an investment of 380 billion yuan in AI infrastructure construction and set a long-term goal to expand the energy consumption scale of its cloud data centers tenfold by 2032. It's evident that Alibaba is firmly committed to AI.

As the market heats up, domestic mainstream AI products like Doubao and DeepSeek, each with around 150 million monthly active users, indicate that C-end market awareness and usage habits for AI products have moved beyond the early education stage and may be entering a mature phase of rapid popularization and in-depth application. At this market inflection point, Alibaba's management believes that large model capabilities have entered the Agentic AI era of 'assisting humans,' and the timing for a large-scale entry into the C-end market is ripe.

During the earnings analyst call, Alibaba Group CEO Wu Yongming clearly stated that Alibaba is making concerted efforts in both AI to Business (AI to B) and AI to Consumer (AI to C) directions.

On the B-end, Alibaba Cloud has established barriers with its infrastructure and industry solutions. According to the second quarter financial results of this fiscal year, Alibaba Cloud's revenue increased by 34%, with external commercialization revenue accelerating by 29%, and AI-related product revenue achieving triple-digit year-on-year growth for nine consecutive quarters.

However, in the C-end market, which directly serves hundreds of millions of consumers, Alibaba has lagged slightly. The QianWen and LingGuang Apps are key moves to address this shortcoming.

In fact, there is immense potential for synergy between AI and Alibaba's e-commerce, maps, local life, and other business ecosystems. Against the backdrop of stagnating internet traffic, Alibaba's core e-commerce and local life businesses urgently need new growth paradigms. AI is no longer just a value-added tool but a core engine for reshaping product experiences, improving transaction efficiency, and unlocking the full lifecycle value of users. Through C-end AI entry points, Alibaba can connect the entire link from demand insight, content generation, and intelligent shopping guidance to after-sales service, building a more robust moat.

On December 9, Alibaba announced the establishment of the QianWen C-End Business Group, headed by Wu Jia, Vice President of Alibaba Group. It is reported that this business group includes businesses such as the QianWen App, Kuake, AI hardware, UC, and Shuqi.

In internal communications, Alibaba mentioned that the primary goal of the QianWen C-End Business Group is to develop the QianWen App into a super app and the first entry point for users in the AI era. In the future, it will further transform the QianWen App into an ubiquitous AI assistant covering scenarios such as glasses, PCs, and automobiles, aiming to 'enable every ordinary person to use AI anytime, anywhere, and continuously benefit from it.'

This ecological integration capability may become a key barrier for Alibaba in future AI competition. Coupled with its technological advantages in cloud computing, big data, and AI algorithms, as well as its deepened open-source strategy, Alibaba may build a more comprehensive AI ecosystem.

Has China's Domestic Large-Scale Model Caught Up?

The phenomenal explosion of Alibaba's QianWen App and LingGuang App is an inevitable result of the technological accumulation and ecological maturity of domestic large-scale models, as well as a microcosm of the collective outbreak of domestic AI.

From a scale and speed perspective, China has undoubtedly entered the global first tier. Data shows that China contributes the highest proportion of globally released large-scale models. Models represented by Alibaba's QianWen series, Baidu's Wenxin, ByteDance's Doubao, Zhipu's GLM, and DeepSeek have demonstrated comprehensive capabilities on par with or even locally superior to the world's top models in various benchmark tests, including Chinese language understanding, mathematical reasoning, and code generation.

According to Stanford University's '2025 Artificial Intelligence Index Report,' by the end of 2024, the performance of top large-scale models from China and abroad has become comparable, with the quantity and quality of China's high-performance AI models improving simultaneously.

More importantly, unlike the closed-source strategies commonly adopted by U.S. giants like OpenAI and Google, leading Chinese companies generally advocate for open-source strategies.

A joint report by the Massachusetts Institute of Technology and the open-source community Hugging Face shows that over the past year, globally downloaded open-source AI models developed in China have accounted for 17.1% of the total, surpassing the United States' 15.8% and ranking first globally, marking that the domestic large-scale model open-source ecosystem has officially entered the global leading ranks.

Alibaba's Qwen series, DeepSeek, and other open-source models have gained widespread application and derivatives in the global developer community, building a vibrant ecosystem. This has significantly lowered the barriers to applying AI technology, enabling small and medium-sized enterprises and individual developers to access core capabilities at low cost, thereby accelerating penetration into various industries.

Therefore, from the perspective of application breadth, domestic large-scale models have demonstrated strong vitality. Their application scenarios have long surpassed early chat-based interactions, penetrating deeply into thousands of industries, including manufacturing, agriculture, consumer electronics, content creation, and enterprise services. This large-scale, diversified scenario refinement also provides massive data and feedback for model iteration, forming an application-driven development path.

Zijin Finance believes that domestic large-scale models have achieved a key transition from 'following' to 'running alongside' global leaders, even taking the lead in some areas.

However, it cannot be denied that at the technological level, there is still room for improvement in basic theoretical research and core algorithm innovation for domestic large-scale models, with some underlying technologies lagging behind international top levels. The AI competition is a long-term race, and current victories do not determine the ultimate battlefield.

Conclusion

For Alibaba, the launch of the QianWen and LingGuang Apps is just the beginning. Whether it can achieve true connectivity of ecological data and services and find irreplaceable user value remains to be seen. This requires not only continuous technological advancement but also tests its organizational strategic judgment and courage.

The competition among domestic AI large-scale models is bound to enter a comprehensive contest in dimensions such as ecological synergy, user experience, and industrial integration. Whether they have caught up will be determined by hundreds of millions of users through their daily usage, providing answers with every real-world interaction.