Tencent and Alibaba are directly competing in the AI Agent space. WeChat is preparing to integrate Agent capabilities within its ecosystem; after the news broke, Tencent’s stock rose over 10% in a single day, with the market assigning a premium of HK$410 billion. Alibaba, meanwhile, is accelerating the opening of its Agent ecosystem—Qwen has fully opened access to third-party Agents, with early adopters including Luckin Coffee, KFC, and China Eastern Airlines currently testing the platform. Qwen now handles over 100 million daily conversations, serves 130 million users using Agents for tasks, and facilitates 300 million transactions, 90% of which originate from Qwen. Their approaches differ: WeChat is layering AI capabilities onto its established social network, while Qwen leverages two decades of accumulated commercial infrastructure as a platform. The article argues that AI Agents will not result in a single winner; instead, Doubao, Qwen, and ChatGPT will each dominate different segments of the AI application landscape.
Article author and source: Wall Street Journal
The capital markets haven't been this excited over a product rumor in a long time.
On June 2, 2026, reports indicated that Tencent was preparing to integrate Agent capabilities within the WeChat ecosystem. Even before the news was officially confirmed, the market had already reacted: on that day, Tencent's stock price rose more than 10% in a single day.
Tencent, which had long been criticized for its slow AI progress and suffered a prolonged decline, has finally regained some investor attention.
A securities analyst specializing in the TMT sector told Wall Street Journal: “The market is rushing to buy because Tencent is adding a task-capable agent on top of WeChat, which has over a billion daily active users and the most dense social network—this could create enough upside potential to drive a revaluation.”
While the WeChat Agent has sparked widespread discussion, Alibaba is also quietly preparing its move.
Wall Street Journal learned that Alibaba is accelerating the opening of its Agent ecosystem. Recently, Qwen has entered high-frequency consumption scenarios such as travel and dining, with companies like China Eastern Airlines, KFC, and Luckin Coffee beginning to participate in Agent collaboration tests. Meanwhile, Alibaba-affiliated services including Fliggy and Taobao are gradually integrating Agent capabilities.
Alibaba has arrived with a familiar story, ready to reimagine its decades-old expertise in matching supply and demand using agents, directly challenging Tencent.
A showdown between top-tier players
The WeChat agent is getting closer to launch.
A person close to the WeChat Agent project revealed that the WeChat Agent was originally scheduled for internal testing in early June, but the WeChat team is considering further refinement and optimization, potentially delaying its final launch. According to the Financial Times, there may be several versions of the WeChat Agent in internal testing.
Outside, Alibaba is quietly making moves, preparing to open the door to its circle of connections.
Over the past six months, Alibaba has been conducting stress tests using its own ecosystem, with dozens of Alibaba-based agents gradually integrating—AI shopping agents embedded in Taobao, instant retail agents in Flash Purchase, travel planning agents in Fliggy, and testing underway on Xianyu.
An agent industry professional told Wall Street View: “Actually, Qwen is China’s largest commercial endpoint for AI-powered trading.” However, he added, “The supply side’s offerings could still be more diverse.”
Zhu Lin's words came true: just before the WeChat Agent's debut, Alibaba struck first.
On June 3, Qwen announced it would fully open its platform to third-party agents and skills, enabling enterprises to operate their own branded agents on Qwen. Leading companies such as Luckin Coffee, KFC, Mixue Ice Cream & Tea, and China Eastern Airlines are currently testing agent services on Qwen and will be launched progressively.
First prototype internally, then expand the ecosystem on the platform—that’s Alibaba’s typical approach.
An executive within the Alibaba ecosystem explained to Wall Street Journal: "An agent fundamentally enhances service quality and addresses customer issues and service efficiency, ultimately winning consumers' choice of this service."
“For example, when using Qwen to book a ride, you simply tell it verbally where you want to go and that your budget is between 10 and 15 yuan, and options appear immediately,” said the individual. “You don’t need to open an app, wait for ages, and manually type in details or fill out forms.”
In his view, Alibaba encourages players to create more agents. “The current focus should be on getting agentic commerce circulating within China’s broader economic environment; otherwise, all commerce will only amount to external coding.”
Wu Jia, President of the Qwen Business Group, told Wall Street View directly, “What we truly want is to integrate AI into the daily life scenarios of ordinary people.” He is confident that this is something that will inevitably happen in the future.
At this point, the most fundamental difference between Tencent and Alibaba lies in how users engage with this new product.
Users on WeChat primarily seek social interaction, communication, and relationship maintenance. All tool functionalities (payments, mini-programs, official accounts) are byproducts of social scenarios.
Users of the Qwen app have a more straightforward purpose: they come to get things done—booking flights, ordering coffee, checking rewards points, or planning itineraries. They don’t use Qwen to pass the time; they come to complete a specific task and leave.
In this mindset, the agent doesn’t need to compete for social attention—it simply needs to execute tasks smoothly. This difference will determine the future business models of the two products.
Tencent's question: Introducing a new character into a familiar social circle
Integrating the Agent with WeChat may sound straightforward, but it might not be as simple as it seems.
The feature density of WeChat ranks among the best in the global internet product landscape. A single chat window supports instant messaging, access to Moments, official account subscriptions, video accounts, mini-programs, payments, enterprise communication, government services... Every new feature added crowds into an interaction path users have already internalized through muscle memory.
And to a large extent, WeChat’s core mindset is “talking to people,” not “talking to systems.” When you open WeChat, the default counterpart on the other end is a specific person—a friend, colleague, family member, or someone familiar from a group.
This mental anchor is too deep—so deep that the presence of the mini-program is suppressed, and most users enter the mini-program only temporarily due to a specific scenario, then leave immediately after use.
To integrate an agent into such a product, Tencent must answer: Why would users turn to an AI within the same window they’re using to talk to friends and say, “Book me a flight”?
Tencent can certainly find ways around this—by creating a standalone AI entry point without altering WeChat’s main chat interface. However, the WeChat Agent would merely become one of many entry points, failing to benefit from WeChat’s true traffic advantages; any change would carry the risk of disrupting users’ established habits.
This is what Tencent truly aims to prove: whether it can create a new mainstream, AI-native interaction within an established, highly saturated national-level product. The challenge lies not in technology, but in product discipline—knowing what to change, what to keep unchanged, and to what extent.
The ultimate direction of the WeChat Agent remains open for now.
It might be a hybrid of social and AI, or it could lean more toward a tool-oriented approach. But no matter the choice, Tencent must confront the usage habits formed by WeChat’s 1.4 billion users—this is both its greatest asset and its greatest burden.
Looking back, Tencent has done similar things before. WeChat Pay originally carved out a financial use case within close-knit social interactions, leveraging the unique catalyst of red envelopes. Similarly, Video Channels entered the already mature short-video market by reversing course through WeChat’s social relationship chain.
Tencent accomplished both of these things, but each took three to five years. For the Agent battle, the market is willing to pay a premium of HK$410 billion upfront; now Tencent must deliver real value in return.
Alibaba's question: Build the business infrastructure for an Agent
For Alibaba, it does not have a national-level messaging product like WeChat. This starting point means Alibaba cannot follow Tencent’s path of “adding AI on top of a familiar social network.”
But Alibaba has its own trump card.
From the perspective of industry insiders, whether an agent can book flights for users depends on whether airlines are willing to open access to their flight systems, seat availability, change and cancellation policies, and special meal preferences, and whether they are willing to allocate resources to operate this workflow. The same applies to coffee brands, food service providers, and government service platforms.
Without this infrastructure work, the agent business model won’t gain traction; large model companies aren’t skilled at collecting payments, merchants don’t trust agents to make decisions on users’ behalf, and users aren’t willing to hand over payment authority.
This is a comprehensive competition involving B2B sales networks, merchant operational systems, and supply chain integration capabilities.
Alibaba has spent over two decades building connections in these areas—from linking foreign trade factories with overseas buyers in the B2B era, to connecting stores with consumers during the Taobao era, and now linking food brands with food delivery users in the local life era.
This inherent genetic trait cannot be grown from scratch by other large model companies.
Not only is the foundation important, but Alibaba has also taken an early step in thinking about demand-side aspects in the Agent era.
A person close to Alibaba gave Wall Street Journal an example: If a user believes they would only buy a product priced at 10,000 yuan once it drops to 7,000 yuan, this demand was previously nearly impossible for merchants to detect. But an agent can place a long-term order on the user’s behalf: “Notify me to buy when it reaches 7,000 yuan.”
As more users express similar demands, what merchants see is no longer vague traffic, but genuine purchasing intent. This is somewhat like a reverse group buy, where the Agent may even help merchants identify demand.
In the past, merchants did not know customers' true price expectations. In the future, agents may directly relay demand feedback to the supply side, giving the business world near-real-time demand signals.
In the age of agents, understanding intent often holds the greatest value. Alibaba is doing exactly that—connecting merchant agents with consumer agents.
Looking back now, Alibaba’s direction and path were quite clear.
Over the past year, Qwen has consistently invested in AI shopping, flash sales, and other initiatives as pilot projects and stress tests. Once these workflows are fully operational, they will be opened to all merchants.
From this perspective, Alibaba is building the underlying network for the Agent era. If this story succeeds, Alibaba’s and Qwen’s valuations could very well undergo a new round of reassessment.
Qwen's current position is somewhat like WeChat mini-programs in 2017—previously, the market viewed it primarily as Alibaba's AI infrastructure investment, in other words, a "money-burning project."
But when daily service conversations exceed one hundred million, 130 million users rely on Agent for tasks, and 90% of 300 million transactions originate from Qwen, it has already met certain conditions for platform valuation.
More notably, the logic of traffic value has changed.
After opening up to third-party agents, Qwen will no longer be just an internal tool product of Alibaba, but will transform into a two-sided B+C platform. This entry point could become a new channel for Alibaba's local life services and e-commerce businesses to acquire incremental users, or it could alternatively reshape the business models of Taobao, Fliggy, and Flash Sales individually.
Of course, this path also carries its own uncertainties: the timeline for widespread adoption of agents, whether Tencent could leverage WeChat to gain a competitive advantage, and whether Alibaba can maintain its own momentum.
One thing is certain: the battle among agents won't be a winner-takes-all story, because Doubao, Qwen, and ChatGPT are taking fundamentally different paths.
Doubao leverages ByteDance’s traffic DNA to excel in companionship and content; ChatGPT leverages OpenAI’s enterprise customer base to sell tools; Qwen leverages Alibaba’s business network to build a platform—these three paths will ultimately occupy different extremes of the AI applications赛道.
However, when an agent can remember your preferences, proactively remind you to place an order before the lunch rush, calculate the optimal redemption strategy when your points are about to expire, and instantly place your order the moment your desired price appears, you’ll realize that the “use-and-leave” user relationships of the past two decades are being replaced by AI assistants that truly understand you.
This replacement process will most likely unfold simultaneously in two entirely different ways, on Alibaba’s and Tencent’s respective home turf.
Tencent and Alibaba are directly competing in the AI Agent space. WeChat is preparing to integrate Agent capabilities within its ecosystem; after the news broke, Tencent’s stock rose over 10% in a single day, with the market assigning a premium of HK$410 billion. Alibaba, meanwhile, is accelerating the opening of its Agent ecosystem—Qwen has fully opened access to third-party Agents, with early adopters including Luckin Coffee, KFC, and China Eastern Airlines currently testing the platform. Qwen now handles over 100 million daily conversations, serves 130 million users using Agents for tasks, and facilitates 300 million transactions, 90% of which originate from Qwen. Their approaches differ: WeChat is layering AI capabilities onto its established social network, while Qwen leverages two decades of accumulated commercial infrastructure as a platform. The article argues that AI Agents will not result in a single winner; instead, Doubao, Qwen, and ChatGPT will each dominate different segments of the AI application landscape.
Source: Wall Street Journal