Text | World Model Workshop
A week ago, Wang Xing said something at Meituan's Q1 2026 earnings call that many people might not have fully grasped.
He said that in the future, Meituan's service targets will no longer just be consumers (To C) and merchants (To B); serving AI Agents (To A) is becoming increasingly important.
To C and To B have been the fundamental business models for the internet over the past two decades.
But now, Wang Xing has proposed a new term: To A.
The truly radical part of this statement is that he is not treating Agents as tools, but redefining them as customers.
If Agents are customers, Meituan must consider how to make Agents more willing to recommend Meituan, not just make users more willing to open the Meituan app.
Behind this lies the quiet replacement of the internet's distribution logic.
And the pace of this replacement is faster than imagined.
The To A Era Business War
In the same week Wang Xing put forward his "To A theory," three other things happened simultaneously.
At Meituan's earnings call, Wang Xing announced that the XiaoMei AI Agent would deeply integrate with Tencent's Yuanbao.
When users say "I want to order takeout" in Yuanbao, Meituan takes over the subsequent steps of choosing a restaurant, placing the order, and delivery.
Just recently, JD.com was also revealed to be in talks with Tencent regarding AI Agent cooperation, and is already connecting with terminal manufacturers like Huawei, OPPO, and Honor.
Across the Pacific, OpenAI announced that ChatGPT will transform into a super-app, integrating external applications like Booking, Spotify, Expedia, etc., focusing on Agents that can autonomously help users complete various tasks.
Without any coordination, it seems like a collective action triggered by the same signal—tech giants are beginning to ally.
This is unusual.
In previous business wars, TikTok links were directly blocked by WeChat, Alibaba and Tencent mutually blocked links for nearly a decade, and Meituan and Ele.me fought a fierce subsidy battle.
The culture of internet business warfare has always been open confrontation, a fight to the death.
But this time, they are allying, and allying quickly. Why?
The reason lies in Wang Xing's statement.
Today, the user's path is: wanting to order takeout, open Meituan; wanting to buy plane tickets, open Ctrip; wanting to shop, open Taobao or JD.com.
Each app guards its own entry point, with traffic circulating within its own pool.
But when Agents enter the picture in the future, users only need to say one phrase: "Help me book Japanese food for tomorrow night." The Agent understands the intent, calls the service, completes the order, without needing to open a single app.
This is the real danger that the big players sense.
If users in the future only talk to Agents, will they still be part of the chain?
Looking again at the intensive alliances between the giants this week, the logic becomes clear.
XiaoMei integrating with Tencent Yuanbao, JD.com partnering with Tencent and bringing in Huawei, OPPO, Honor—these are essentially moves to "To A," to secure recommendation slots within Agents.
So this is not an ordinary business war. With the arrival of the To A era, the internet is undergoing a re-division of labor, and all players want to defend their territory.
Different Paths to To A
Notably, different companies are approaching To A differently. Currently, this wave of alliances is taking three paths:
First type: Super entry point + Service providers.
Tencent Yuanbao, WeChat, ChatGPT are all doing this.
Tencent Yuanbao integrating Meituan, Tencent's traffic entry point integrating JD.com, OpenAI bringing Booking, Spotify, Expedia, Canva, etc., into ChatGPT—essentially, these are packaging services like takeout, shopping, travel, content, design, payment into a single Agent entry point.
They are competing for the user's first stop when making a request.
Second type: Apps packaging themselves into callable services.
Meituan XiaoMei, JD.com AI Agent, Taobao, Uber, Expedia, OpenTable belong to this category.
Their logic is pragmatic: if users no longer open us directly in the future, then I must at least ensure that when the Agent makes decisions for users, I can still be called.
Better to retreat from the front-end entry point to a back-end capability layer than to be completely bypassed.
Third type: Smartphone manufacturers building system-level Agent entry points.
Huawei's Xiaoyi, Honor's YOYO, OPPO, Xiaomi are taking a more fundamental route.
They may not necessarily build their own takeout, shopping, or social apps, but they control the smartphone system's entry point.
The user's first spoken request might be heard first by the phone's AI assistant, then distributed to WeChat, JD.com, Meituan.
This is an opportunity for smartphone manufacturers, who lost the entry point in the App era, to re-position themselves using Agents.
However, some companies have taken a completely different path.
For example, Alibaba has chosen to first integrate internally.
Qianwen, Taobao AI shopping guide, Alipay AI wallet, DingTalk AI travel all connect Alibaba's own services. Fliggy, Amap, Taobao Quick Buy—all are within the integration scope.
Alibaba is choosing to first turn itself into a complete closed loop, so when it exports externally, it's already a packaged Alibaba service layer.
Regardless of the path, what everyone is competing for is an irreplaceable position in the new To A chain.
Where Will To A Lead?
Alliance is the first step in this business war, but not the end.
The current landscape seems like a win-win:
Tencent Yuanbao gains Meituan and JD.com's service capabilities; Meituan and JD.com gain WeChat's entry point traffic; Huawei, OPPO, and other smartphone manufacturers get on board, each getting what they want.
But there is a natural crack in these collaborations: the interests of the entry point holder and the service provider have never been completely aligned.
Could Yuanbao, which calls Meituan today, directly allow restaurants to register on its own platform tomorrow, skipping the Meituan layer?
Could ChatGPT, which integrates Booking today, connect directly to hotel inventory tomorrow, no longer needing the OTA as a middleman?
This is not a conspiracy theory; it's a commercial temptation any platform that controls a sufficiently large entry point would face.
When WeChat launched Mini Programs, countless lightweight apps were absorbed; when Google Maps went online, local navigation apps collectively disappeared.
This time, if a super-app could truly enable suppliers to be directly dispatched by its own Agent—restaurants, hotels, ride-hailing, shopping, all bypassing existing aggregation platforms—then the existence of apps like Meituan and Ctrip would be in danger.
Although this path is difficult, it's not entirely impossible.
Therefore, today's alliances among giants are essentially service providers securing their spot before the window closes, preferring to be called than to be bypassed.
They are betting that their service capabilities are solid and hard to replace, making the Agent entry point dependent on them.
Beyond this, there are other questions in this To A battle.
For example, could Agent service recommendations eventually evolve into a new form of paid placement advertising?
If so, Meituan and JD.com would have to pay for the user traffic they originally owned, adding an extra layer of toll.
Another example: if the Agent's recommendation results have problems, who is responsible? The entry point holder or the service provider?
No one knows the answers, but everyone has already started running.
After all, in this To A reconstruction, the greatest danger is not losing the race, but not hearing the starting gun.
