The real battlefield isn't about "who is smarter," but about "who can get AI to spend money, operate accounts, and make decisions on behalf of users."

Author and source: 0x9999in1, ME News



TL;DR

• On June 15, 2026, The STAR Market Daily exclusively reported: Ant Group is currently testing an AI-powered version of Alipay, featuring an AI assistant named "Abao" that supports direct voice or text commands to book rides, order coffee, and purchase funds; with authorization, it can also execute investment operations.
• On Tencent’s platform, the prototype testing of WeChat’s AI agent has been completed, with compliance approval expected to begin as early as June and a full public rollout planned for Q3. On the day the news broke, Tencent’s stock rose nearly 10% intraday, adding approximately RMB 360 billion to its market capitalization in a single day.
• User base: WeChat has approximately 1.432 billion monthly active users; Alipay has 969 million. This is not a competition between two apps—it’s a national-level battle between two ecosystem gateways.
• Ant's strengths: AI Payments surpassing 300 million transactions in three months, Lingguang app, and Alipay reaching over 30 million monthly active users. Tencent's strengths: WeChat with 1.2 billion monthly active users, millions of mini-programs, and newly launched AI ecosystem integration capabilities.
• The real battlefield isn't about "who is smarter," but about "who can let AI spend money, manage accounts, and make decisions on behalf of users"—this is Alipay's natural stronghold.
• But Bao faces a dilemma: if the selling point is simply using voice to hail a ride or order coffee, it’s directly competing with WeChat’s strengths.
• The key to winning this battle isn't the model—it's the entry point, trust, and organizational collaboration.

I. Who is Abao, and why does he deserve a headline?

First, state the facts.

On the morning of June 15, 2026, The STAR Market Daily exclusively reported: Ant Group is currently in internal testing an AI-powered version of Alipay. The new version revolutionizes the existing interface by allowing users to instantly switch to a native AI interface. The AI assistant’s Chinese name is tentatively set as "Abao."

What can Abao do? You can hail a ride, order coffee, place food deliveries, or check market trends. After authorization, you can also buy funds and manage your investment accounts. It works with both text and voice commands.

What does this mean?

It means Alipay is bidding farewell to the "function menu + search bar" model. It means users no longer need to navigate through a dozen subpages to find what they need. It means the logic of "people seeking services" has been replaced by the logic of "services seeking people."

Going further—it means "people don't even need to look, they just need to say."

Ant Group has not responded to this matter. But the version is on its way, and the window of opportunity has opened. This is not a buildup—it’s a move.

Why now?

Because the card on the other side is about to be flipped too.

Second, Tencent's card has been held for over a year.

Shift your focus to Shenzhen.

On June 2, 2026, when Caijing asked Tencent about the progress of WeChat's AI agent, the response was unusually cautious: "The launch timeline cannot be determined at this time." However, media outlets have obtained leads: prototype testing has been completed, and compliance review could begin as early as this month.

The capital market was even more excited than Tencent itself. On the day the news broke, Tencent's stock price surged nearly 10% intraday, adding approximately HK$360 billion to its market capitalization in a single day.

$360 billion. The anticipated AI transformation of a super app can mobilize this scale. This isn't a bubble—it's the market's genuine valuation of the "super gateway + AI" combination.

More intriguing is June 8, when the WeChat Open Platform officially announced the opening of WeChat AI ecosystem integration capabilities to all mini-program developers. Two modes are available—“Automatic Mode,” where the platform automatically adapts by reading the source code, and “Development Mode,” where developers can deeply customize their own integrations.

An industry saying: This is paving the way for WeChat's AI agent.

How wide a road to build? Millions of mini-programs. This means that in the future, when a user tells an agent in WeChat, “Help me book a high-speed train ticket,” the request may not be handled by Tencent’s own services, but by any authorized mini-program.

This has been a closely guarded project by Tencent for over a year. According to multiple media reports, the project was initiated in the first half of 2025 and classified by Tencent as a "top strategic priority," with the internal codename "Top Secret." The original plan was to conduct a limited beta test in mid-2026 and open it to all users in Q3.

Note the four characters: "all users".

WeChat's monthly active users, according to QuestMobile data as of March 2026, have reached 1.432 billion, making it one of the largest active user bases on Earth.

III. Why this battle must be fought and cannot be lost

Who is more eager, Ant or Tencent?

If we look only at public sentiment, Tencent is more eager. Ma Huateng himself said at last year’s shareholder meeting: “A year ago, we thought we had boarded the ship, but then we found it was leaking, so we started looking for another one. Now we feel we’ve stepped onto it, but we still can’t sit down comfortably—we still hope the ship can move faster.”

This analogy is very Tencent—no pretense, no feigned composure, just laying the anxiety bare for the market to see.

Why be anxious? The data is right there.

DouBao reached 100 million daily active users and 226 million monthly active users by the end of 2025. DeepSeek achieved 135 million monthly active users. Even with the momentum from Tencent’s 2026 Spring Festival red packet campaign, Tianjin Yuanbao reached only 114 million monthly active users—less than a fraction of DouBao’s numbers. In this foundational model battle, Tencent has already lost the first phase.

Even more awkward was the product timeline. At the end of 2024, the Yuanbao and Hunyuan large models were only officially transferred under CSIG for product development. By March 2026, Tencent shut down its AI Lab, which had operated for a full decade, and integrated its personnel into the Hunyuan system.

Money has also been burned. In Q1 2026, Tencent’s quarterly capital expenditure reached RMB 31.936 billion, a 16.23% year-over-year increase and a record high. R&D spending was RMB 22.54 billion, up 19% year-over-year. For four consecutive quarters, “payment volume has exceeded recognized volume”—in plain terms, this means paying premium prices to secure computing power.

But here’s the problem: money can buy computing power, but it can’t buy user mindshare. Models can be caught up to, but organizational alignment cannot be recovered.

WeChat has become Tencent’s last surefire ace in the hole.

What about Ant? On the surface, they seem calm, but in reality, the pace is even tighter.

Ant Group does have AI products. In November 2025, it launched its general-purpose AI assistant, "Lingguang App," featuring "generate an app in 30 seconds." In April 2026, it intensified efforts with the "Lingguang Flash App Creator Incentive Program," allocating a dedicated fund of 100 million yuan. As of January 2026, the health AI app "Ant Afu" had surpassed 30 million monthly active users.

But these are all side quests.

Where is the main line? It's within Alipay itself.

On May 26, 2026, Ant Group announced the completion of its AI payment ecosystem rollout and launched the world’s first Token Pay service, supporting 95% of general-purpose agents. As of May, the number of transactions via Alipay AI Pay exceeded 300 million within three months.

300 million transactions. This isn't a demo—it's real-money trading activity.

But is 300 million enough?

Not enough. This is just "AI making Alipay payments from outside." The real story is the opposite—"AI making decisions and accessing your account within Alipay." That’s the problem Bao wants to solve.

Fourth, on the surface, it's about competing for users; at its core, it's about competing for entry-point paradigms.

Many people interpret this battle as "two apps competing for users." This understanding is too superficial.

It’s not about acquiring more users. WeChat and Alipay have already surpassed 1.4 billion and 969 million monthly active users, respectively. Where is the growth coming from? There is no more growth.

What are you rushing for?

Entry point paradigm.

What has been the core entry logic of China's internet over the past 20 years? Icons. The 3x3 grid. The search bar. "You open an app and tap a button."

After the AI agent arrived, the logic changed—“Say one thing, and it’s done.”

Whoever defines the paradigm of "say one sentence" will rewrite the entry rules for the next-generation internet.

The Alibaba Qwen app has integrated dozens of ecosystem agents, handling over a hundred million service-related conversations daily. ByteDance's DouBao has connected with TikTok Shop, enabling direct purchases. These are all early experiments, but none constitute a direct showdown in the "super app" arena.

Only Alipay and WeChat are true "national-level super apps." Only these two platforms enable AI agents to face, for the first time, a real battlefield of over one billion active users, covering all aspects of daily life—food, clothing, housing, and transportation—and connected to a transaction system worth tens of trillions.

The first party to successfully complete the process will do three things.

First, define users' expectations for AI agents. Once Bao gets users accustomed to "buying funds via voice," WeChat’s subsequent efforts will simply be following suit.

Second, attract the developer ecosystem. Whoever has the largest AI traffic entry point—whether it’s mini-programs, merchants, or service providers—will secure the highest-quality offerings.

Third, reimagine the monetization pathways. Advertising has become "AI recommendations," e-commerce has become "Agent direct procurement," and finance has become "AI advisors proactively making suggestions"—all revenue models must be rewritten.

Any one of these three things is enough to put the other side on the defensive.

Five, what advantages does Abao have, and what are the shortcomings of WeChat agents?

Lay out the cards on both sides to see them.

Bao's greatest advantage is the integration of "account, funds, and credit" into one system.

Alipay is fundamentally a financial account system. While WeChat can also handle tasks like hailing rides, ordering coffee, or buying takeout, Alipay’s core expertise—buying funds, managing investments, adjusting limits, and making credit decisions—has been honed over 20 years.

After authorization, Abao can directly execute investment operations. This statement carries immense weight—it means AI is for the first time truly "touching money," not just "touching services."

This is the part that WeChat’s intelligent agent finds hard to replicate. Although WeChat’s Tenpay also handles payments, the depth in wealth management, funds, and insurance is structurally inferior. Ant Group controls Huabei, Jiebei, and Sesame Credit—all of which are financial infrastructure directly accessible to AI-driven decision-making.

But Bao also has Bao's weaknesses.

If Aobao focuses on scenarios like ride-hailing, ordering coffee, or food delivery, it’s essentially challenging WeChat on its home turf with its own non-core strengths. WeChat boasts 1.432 billion monthly active users and millions of mini-programs, with a far greater density of lifestyle services than Alipay. As 36Kr accurately pointed out, this is using insufficient ecosystem resources to compete against the opponent’s strongest area.

The advantages of WeChat's intelligent agent lie in its ecosystem scale and social connectivity.

Millions of mini-programs mean an extremely rich supply. Social relationship chains enable AI to understand real-life contexts like "Send a red envelope to Old Zhang." Video Channels, Official Accounts, and Enterprise WeChat—all content, information, and enterprise services—are fully integrated within a single app.

What are the weaknesses of WeChat's intelligent agent?

First, Tencent’s proprietary model has not yet gained a leading edge. The HunYuan Hy3 preview was only released in April, with a total of 295 billion parameters, still lagging behind top overseas models. Tencent plans to invest RMB 18 billion in AI in 2025, while ByteDance will invest RMB 80 billion during the same period—a more than fourfold difference that cannot be closed within just a few quarters.

Second, the WeChat agent does not read chat records. This is a compliance boundary repeatedly emphasized by Tencent. However, this also means it cannot make personalized decisions based on the user’s most authentic context.

Third, Tencent’s AI business is eating into profits. The Q1 2026 financial report showed that the new AI business alone reduced operating profit by approximately RMB 8.8 billion in a single quarter. Excluding AI, Non-IFRS operating profit growth was 17%; including AI, it dropped to just 9%. This cost is substantial, but it is necessary.

Neither side is easy.

Six: Compliance, Security, and Trust—This Is Where the Real Moat Lies

Many articles focus only on technology or ecosystems, but overlook the most important issue—can regulators approve AI spending on users' behalf and managing their accounts? And do users dare to allow it?

As of the end of April 2026, 868 generative AI services have completed registration. As a national-level entry point, WeChat faces even greater complexity in compliance approval. Tencent has also confirmed that the compliance approval process will "begin as early as June." How long this process will take remains uncertain.

For Alipay, the compliance threshold will only be higher—because it involves wealth management, funds, and credit. Any financial operation autonomously executed by AI could trigger regulatory inquiries: Was the user’s authorization genuine? Is the AI’s decision-making process traceable? Were risk disclosures adequate?

But conversely, this is precisely Ant's moat.

Ant's capabilities in financial compliance are something no internet giant can easily replicate in the short term. AI Pay processing 300 million transactions in three months without incident is itself proof of this.

Transferring user trust is more difficult.

It’s low barrier to let users hail a ride via voice command on WeChat; it’s ten times harder to get users to authorize AI to buy $10,000 in funds. Getting the first wrong might cost at most $20 extra; getting the second wrong could lead to thousands in losses.

Whoever can build a trust curve around "AI making financial decisions for me" over 6-month, 12-month, and 24-month timeframes will win the second half.

This is not a technical issue. It is a comprehensive effort involving product design, compliance, user education, and risk pricing.

Seven: Although it appears to be a showdown between two giants, a third party is waiting for an opportunity.

Pull back the camera a bit.

ByteDance's Doubao has 226 million monthly active users and has successfully translated its traffic advantage into user mindshare. Alibaba's Qwen is integrated with dozens of ecosystem agents, serving over a hundred million conversational requests daily. Kuaishou's Keling AI has 60 million global users in the video generation space, with revenue of approximately RMB 1.04 billion in 2025, monthly revenue exceeding $200 million in December, a valuation of $20 billion, and plans for an IPO in 2027.

These are not spectators.

If Alipay and WeChat exhaust each other and move too slowly in their competition over AI entry points, ByteDance’s Doubao could very well overtake them with its pure "AI-native app" form. Alibaba’s Qwen is also accelerating its integration into more ecosystems, and Ant’s internal “Lingguang” is another significant but under-the-radar factor.

The landscape of China's AI application layer has not yet been finalized.

According to QuestMobile data, the combined monthly active users of the top five AI apps in Q4 2025 reached 455 million: Doubao, DeepSeek, Yuanbao, Afu, and Qwen. Note—Afu, Alipay’s health assistant, has now entered the top five.

Ant's AI strategy has never been about "waiting for Alipay to make the first move." It’s a multi-front campaign with breakthroughs across multiple fronts. Abao is the main force, while Lingguang and Afu serve as the flanks.

Tencent, on the other hand, has gone all in. After monthly active users for Yuanbao failed to grow, the entire AI breakthrough strategy has been pinned on WeChat's intelligent agent alone.

Whose strategy is more robust? Whose bet is more risky? The market will provide the answer, but not immediately.

Eight, a few final words

Back to the most fundamental question: Who will win, Abao or WeChat's AI agent?

My answer is—this is the wrong question to ask.

The winner is not a single company, but the paradigm of "Super App + AI Agent." The two companies are merely two implementers of this paradigm. If one succeeds first, the other must follow. Either you survive, or you die—there is no middle ground.

For users, this is a good thing. After a decade of relying on grid layouts, search bars, and secondary pages, all of these are finally being replaced by a single sentence. Ordering a ride, making a payment, buying funds, checking your balance—you no longer need to remember where each function is located.

But even good things come at a cost.

As AI makes more and more decisions for you, will you still remember your own preferences? When algorithms choose your funds for you, will you truly understand your asset allocation? When everything is reduced to "just say one thing," will your judgment be simplified along with it?

These questions have no standard answer.

On that day, Tencent's stock rose 10%, making the market sentiment clear—the AI transformation of the super app is the most certain narrative. Ant's timing in releasing information about Abao could not have been more precise—it seems anything but coincidental.

The cards are on the table. Now it’s just a matter of who flips first.

Also see who can withstand all the risks after the flip.

After all, the real key to victory in the age of AI agents has never been "who is smarter," but "who can maintain user trust over a sufficiently long period."

The wind blows in whichever direction fate decrees. But one thing is certain—

The old entry paradigm will not return.

Source:

1. Exclusive! Alipay Beta Testing the "Super" App, 21st Century Business Herald / Sci-Tech Daily, June 15, 2026
2. Are WeChat Agents Coming? WeChat Announces That Mini Program Developers Can Integrate into WeChat’s AI Ecosystem, Tencent News / Southern Metropolis Daily, June 8, 2026
3. Will Tencent’s AI Comeback Succeed with the Launch of WeChat AI Agent? GPLP Tech Talk / NetEase, June 11, 2026
4. Big News! AI Alipay Is Coming, Intensifying the Battle for AI Service Entry Points, 36Kr, June 14, 2026
5. Alipay AI Payments Exceed 300 Million Transactions in Three Months: A Deep Dive into Ecosystem Development, 21st Century Business Daily, May 27, 2026
6. QuestMobile 2025 Comprehensive Ecosystem Traffic Annual Report, QuestMobile, February 3, 2026
7. Report: Tencent Is Developing a "Top Secret" WeChat AI Agent, Set to Enter Beta Testing Mid-Year, Sina Finance/The Information, March 11, 2026
8. "Doubao Becomes China's Largest AI Application in 2025, Qwen Experiences Rapid User Growth in Q4," Jingji Ribao, February 2, 2026