Over the past two years, as domestic AI applications have boomed and intensified in competition, Alibaba, with its focus on open-source models, has maintained a relatively subdued presence in the consumer (C) market.

However, the landscape is now shifting.

On November 17th, Alibaba officially unveiled its 'Qianwen' initiative, launching a public beta version of its app for widespread use. By the 19th, the Qianwen app had soared to third place on the App Store's overall free app chart. After just one week of public testing, downloads had surpassed 10 million. Alibaba hasn't staked its entire future on Qianwen alone; instead, it has adopted a multi-pronged strategy. Alongside the Qianwen app, the all-modal general AI assistant 'Lingguang' was introduced. Subsequently, Kuake and Qianwen underwent deep integration and were jointly relaunched.

The sequential launches of Qianwen, Lingguang, and Kuake signal Alibaba's gradual expansion in the AI-to-Consumer (To C) arena, with the ultimate ambition of becoming users' primary gateway in the AI era.

Competing for gateways equates to competing for traffic and users. Given that AI applications can infiltrate nearly all internet transaction scenarios, they hold immense appeal for industry giants.

Anticipated changes include a shift from traditional user searches to AI-driven decision-making, potentially disrupting advertising and transaction models. For instance, ChatGPT's collaboration with e-commerce enables users to complete purchases within the chat interface, while Doubao exclusively features shopping links from Douyin Mall.

However, numerous early entrants have already staked their claim to this gateway position: Doubao leads the pack, with Kimi, Wenxin, Yuanbao, and others in close pursuit.

Alibaba possesses the determination, financial resources, and technological prowess, but the gateway competition will not be easily won. Moreover, it must confront the reality that in competing with 'newcomers,' it has transitioned from offense to defense and is on the brink of a protracted, all-out war.

Venturing into the application layer, Alibaba's first formidable opponent is ByteDance.

Before Qianwen, Alibaba had concentrated its C-end resources on Kuake. However, Kuake's underlying logic, rooted in browsers and searches, has made it challenging to alter users' perception of it as a mere tool, thereby impeding its development in the AI application space.

A 'useful browser and tool' clearly cannot fulfill the vision of a super gateway.

Now, with Qianwen's launch as an independent application, unburdened by historical constraints, it is better positioned to realize Alibaba's strategic aspirations in the To C market.

Nevertheless, it's undeniable that from Kuake to Qianwen, Alibaba seems to have missed a crucial window, rendering it a latecomer in the AI gateway race.

Whether it can stage a comeback is a question on everyone's mind. Although initial days have witnessed high popularity, long-term success is not guaranteed. After all, this is a common highlight moment for many domestic AI launches.

Without the first-mover advantage, sustaining an upward growth trajectory is an extremely complex and challenging task for Alibaba.

Of course, Alibaba's accumulation in underlying large model capabilities provides Qianwen with a robust technical foundation, which may be its greatest confidence in challenging the AI application market.

The Qianwen app leverages the Qwen series models and integrates Tongyi Lab's most potent Qwen3-Max. Since going fully open-source in 2023, Alibaba's Qwen models have amassed over 600 million global downloads, with over 300 open-source models and over 170,000 derivative models, surpassing Meta's Llama to become the world's premier open-source model family. Currently, Qwen models serve as the technical cornerstone for numerous Silicon Valley startups.

The recent release of the trillion-parameter model Qwen3-Max further solidified Alibaba's technical support for underlying models. In early November, Qwen3-Max-Thinking released a preview version, achieving perfect scores in mathematical competition tests like AIME 25.

Prioritizing models before products, Alibaba has chosen a relatively stable path in the AI era, which is also reflective of its identity as a 'full-stack' player. Like Google, Alibaba is among the few global internet giants that have constructed a comprehensive ecosystem, spanning from underlying chip hardware and middle-layer model architectures to upper-layer applications. Building such full-stack technical capabilities often necessitates long-term technical accumulation and massive investment.

With a solid foundation, creating 'gateways' becomes more feasible. Relatively, the stronger the underlying large model capabilities, the greater the opportunity to attract users at the application layer.

Beyond technology, Alibaba's vast ecosystem provides Qianwen with a wealth of scenarios, facilitating its transition from conversation to action. 'Being able to take action' may well become Alibaba's differentiated feature in the C-end market.

Integrating technological, business, and ecological advantages, Alibaba has the potential to create a mature AI product. However, in the complex and fierce domestic AI application landscape, surpassing leading competitors will not be easy. Data indicates that in the domestic AI application market, the top two players have achieved a dominant leading position, especially Doubao, which has successfully overtaken Deepseek and continues to grow strongly.

Facing Doubao, which enjoys a first-mover advantage and is unstoppable, Alibaba's AI products resemble a 'newcomer,' much like when Douyin first entered e-commerce and faced Alibaba, full of fate and cyclical significance.

According to QuestMobile's Q3 2025 AI application industry report, Doubao under ByteDance topped the chart with 172 million monthly active users, surpassing Deepseek's 144 million. Tencent Yuanbao ranked third but with 32.86 million monthly active users, far behind the top two.

As for Kimi, Zhipu Qingyan, and MiniMax, once part of the 'AI Four Little Dragons,' they all experienced a decline this quarter, with Kimi falling below the 10 million mark.

Currently, ByteDance is undoubtedly the most promising contender, boasting not only Doubao but also rapid user growth for Jimeng. Importantly, ByteDance's ambition in the AI era is not just to become a new traffic pool but to form a complete closed loop where search, shopping, and consumption services can be achieved without opening other apps or browsers.

Allowing one of China's largest traffic pools to become a new super gateway and turning traditional e-commerce into mere 'shelves' could become Alibaba's nightmare.

Moreover, Douyin E-commerce is thriving, with its payment GMV in the first ten months of this year approaching Pinduoduo's, and its advertising revenue last year surpassing Alibaba's.

On one hand, there's the new gateway competition in AI; on the other, there's the fierce battle in e-commerce's core business. Thus, Alibaba and ByteDance are about to engage in an all-out war. However, this time, ByteDance resembles the stronger 'defender.'

Finding Scenarios with Technology vs. Developing Technology within Scenarios

The emergence of Qianwen is not just about competing with other giants for traffic gateways in the AI era; it also embodies Alibaba's aspiration to build a 'unified' super gateway.

As the 'central nervous system' of Alibaba's full-stack ecosystem in the future, it has the potential to connect core scenarios within Alibaba's ecosystem, such as payments, travel, shopping, and entertainment, breaking down barriers between business segments and enhancing the synergy of Alibaba's various ecosystems.

For users, this represents a transformation in experience. For example, if you plan a weekend family trip, booking tickets, hotels, and paying with Yu'ebao, the system can compare prices and place orders on Taobao or automatically complete cross-platform coordination to choose the right plan for you.

Optimistically, once Alibaba controls the gateway, it controls traffic and discourse power. Coupled with AI's strong penetrating power, it could indeed link multiple user consumption and living scenarios, precisely matching needs with services, and maximizing the value of traffic. Ultimately, it could achieve growth breakthroughs that are difficult for single business lines, bringing new vitality to Alibaba.

However, the issue is that this is a typical approach of 'finding scenarios with technology,' which incurs higher costs than ByteDance's approach.

On one hand, when it comes to combining AI and e-commerce, Doubao + Douyin E-commerce vs. Qianwen + Taobao & Tmall seem to have consistent links and logic, with front-end gateways and back-end supply chains and inventory, creating an ideal ecological closed loop when internally coordinated.

But the problem lies in the fact that ByteDance has China's largest native traffic pool, while Alibaba does not. Doubao and Douyin E-commerce both rely on Douyin's super national app, with business and scenarios naturally extending. However, for Alibaba to complete this closed loop, it must first turn 'Qianwen' into a traffic source, which is not achievable solely with financial resources, technology, and determination.

On the other hand, tracing back to Chinese internet product innovations, 'finding scenarios with technology' or having technology but lacking scenarios has proven difficult to overcome user habit barriers and truly launch products. This is a consensus verified by countless internet entrepreneurs.

Alibaba has repeatedly attempted to build super gateways in the mobile internet era, especially focusing on social media, launching Laiwang, DingTalk, and Alipay Social, among others. However, except for DingTalk, almost all have failed. This cannot be entirely attributed to Alibaba lacking a social gene, as both Alipay and Taobao possess massive traffic pools. However, the lack of scenarios is a core reason for Alibaba's unsuccessful attempts.

In the content e-commerce field, Alibaba's resource investment in content ecosystems has not yielded significant growth. Instead, Douyin and Xiaohongshu have successfully become important platforms for user inspiration and consumption, further demonstrating that the lack of scenarios cannot efficiently meet large-scale user demands or change user mindsets.

In the AI era, whether an application becomes a hit is often not solely determined by technological factors; technology, scenarios, and innovation are all indispensable.

From a product strategy perspective, Alibaba resembles a 'Three Heroes Battle Against Doubao,' with Qianwen, Kuake, and Lingguang competing separately, potentially giving rise to a high-frequency assistant tool application.

However, 'universal gateways' and 'closed-loop transactions' may have potential conflicts. The more public a gateway is, the more open it must be; but the more transaction-closed it is, the more barriers it must build. One could even say that open gateways and commercial closed loops may be inherently mutually exclusive. Of course, this is also a consideration for other AI applications seeking to monetize through the consumption chain.

Compared to Alibaba's approach of 'finding traffic with AI' and ultimately feeding back into e-commerce, ByteDance's logic leans more towards 'developing AI within scenarios.' Based on core scenarios of short videos and social media, it cultivates and creates AI applications, continuously updating and upgrading them within these scenarios.

For example, Doubao, within Douyin, meets users' needs for browsing videos, leading to AI summaries, AI bullet comments, and private message AI avatars. For creators, with content creation needs, it generates short video scripts, titles, and footages from keywords, as well as artistic video covers from user-uploaded photos.

Thus, Doubao's functionality becomes increasingly powerful, and the experience relatively rich. As long as it has the potential to become the next super app, capturing users' time and attention, it is a valuable asset.

Can the AI Gateway Resolve Alibaba's Traffic Anxiety?

Finding a new traffic gateway has been Alibaba's unfulfilled dream for over a decade.

After all, almost all businesses based on e-commerce consume traffic and struggle to generate new organic traffic. Users 'buy and leave,' with low usage duration, determined by inherent genetics.

Thus, from 'Laiwang PK WeChat' to 'Qianwen PK Doubao,' the pattern continues.

Moreover, in the AI era, the intensity of gateway competition will be higher. If AI applications bypass e-commerce platforms to complete transactions directly, e-commerce becomes mere inventory suppliers, affecting traditional platform traffic transactions and advertising revenue.

This explains Alibaba's strong determination and expectations for the AI To C gateway. In the mobile internet era, despite controlling the largest e-commerce transaction scenarios, it has failed to create a national-level app that can compete with WeChat and Douyin or a primary traffic gateway that occupies users' time and has global influence.

This may currently become a disadvantage for Alibaba in competing with other giants in the AI application landscape.

In recent years, in the AI application market, most internet giants and startups have chosen to accelerate penetration to C-end users through large-scale investment, seizing favorable positions. Especially the former, with Doubao leading and Yuanbao following, their rapid rise confirms that current AI application competition still relies on traffic acquisition.

As ByteDance and Tencent, China's two largest internet traffic pools, advance deeper into e-commerce and AI tracks, traffic becomes increasingly scarce, as existing traffic pools gradually turn into 'private reserves,' prioritizing internal ecosystems and conversions. This also means that as Alibaba's competitive relationship with ByteDance and Tencent in e-commerce and AI inevitably intensifies, it will become increasingly difficult to purchase high-quality users and traffic from other giants in the future.

However, when we take a broader view of the global market, Alibaba's standing as a leader in open - source models could potentially channel more overseas traffic to its AI applications. This is a distinct advantage that Doubao and Yuanbao simply do not possess.

The AI - to - consumer (AI To C) war is a battle that Alibaba cannot afford to lose; it must not only engage in but also emerge victorious. However, this conflict may well escalate into a prolonged, all - encompassing war that could eventually become a significant drain on its cash flow.

From a financial perspective, both ByteDance and Alibaba boast substantial resources, along with a strong resolve and willingness to invest heavily in AI.

Alibaba had previously unveiled its plans to pour 380 billion yuan into AI and related cloud computing sectors over the next three years, with the possibility of even greater investment in the future.

ByteDance's investment in this field is nothing short of astonishing. According to securities research reports, ByteDance's research and development (R&D) investment far outstrips that of its peers. In 2024, its capital expenditures reached 80 billion yuan, nearly matching the combined total of Baidu, Alibaba, and Tencent (around 100 billion yuan). Some analysts even forecast that this figure will double to 160 billion yuan in 2025.

However, the crux of the matter lies in the fact that Alibaba is currently embroiled in two major battles simultaneously: one in the food delivery sector and the other in AI. Engaging in two high - stakes conflicts with substantial investments, even for a financially robust behemoth like Alibaba, the pressure is mounting significantly.

According to Alibaba's financial report for the second quarter of fiscal year 2026, the group's operating profit took a nosedive, plummeting 85% year - on - year to 5.365 billion yuan. Meanwhile, its adjusted EBITA (Earnings Before Interest, Taxes, and Amortization) also saw a sharp decline of 78% year - on - year, dropping to 9.073 billion yuan. This decline can be largely attributed to its food delivery business. According to estimates by Dolphin Research, Shanguo (Alibaba's food delivery platform) incurred a net loss of approximately 36 billion yuan, which consumed nearly 80% of the profits from its e - commerce segment.

Moreover, Alibaba's cash reserves have been on a continuous downward trajectory and were surpassed by Pinduoduo this quarter. Data shows that as of September 30, 2025, Pinduoduo's cash, cash equivalents, and short - term investments stood at 423.8 billion yuan, while Alibaba's cash reserves had dwindled to 373.6 billion yuan.

In stark contrast, ByteDance is operating with relative ease. With cash reserves exceeding US$50 billion and an asset - liability ratio below 20%, it is able to fund its AI investments and stock repurchases through its own profits, without having to rely on external financing.

Furthermore, Alibaba is currently facing a multitude of competitors beyond just ByteDance. In its core e - commerce business, it has to constantly fend off challenges from Pinduoduo, Douyin, and JD.com. Its foray into the food delivery sector has turned into a protracted war of attrition with Meituan and JD.com. And in the AI arena, it is surrounded by formidable rivals such as Tencent, ByteDance, and Baidu. The intensity of this multi - front competition is likely unprecedented.

New variables may already be simmering beneath the surface, ready to disrupt the current so - called market landscape at any given moment. No company can afford to be complacent. Especially for Alibaba, this is set to be an exceptionally tough battle.

Dao Zong You Li, previously known as Wai Dao Dao, is a new media outlet that focuses on the internet and tech circles. This is an original article. Any form of reprinting without retaining the author's relevant information is strictly prohibited.