Preface

TrueView

The large model industry has long struggled with practical implementation, and ride-hailing happens to be the perfect, lightweight, high-frequency, and tangible enlightenment tool.

Yet, more ruthless than enlightenment is how AI ride-hailing is rewriting the internet's most fundamental power dynamics—shifting from users seeking services to intentions preceding services, and from service ownership to the authority to allocate services on behalf of users.

History does not simply repeat itself, but it often rhymes.

This article will focus on analyzing:

1. Why is ride-hailing the optimal entry point for AI mindset enlightenment?

2. How do the strategic ambitions of Didi, Qianwen, and Doubao differ?

3. Why is the essence of AI ride-hailing not a mobility revolution?

4. How will data and transaction interception reshape profit distribution?

5. Has the decisive logic of this entry-point battle changed, shifting from subsidy wars to infiltration strategies?

Content/Jin Huan

Editor/Yong E

Proofreader/Mang Fu

The so-called disruptive new trends of the era often merely involve remeasuring old territories with new technologies.

In 2014, Didi and Kuaidi ignited a subsidy war on the streets of Beijing, reshaping not only how Chinese people traveled but also sparking the most significant payment entry-point enlightenment movement in China's mobile internet history.

Twelve years later, a similar script unfolds in the same setting. In just a few months, Didi, Qianwen, and Doubao have heavily invested in "AI ride-hailing" features.

On the surface, this appears to be an underwhelming interaction upgrade—users simply voice their needs, and the system automatically handles intent dissection and order dispatch.

However, beneath the homogeneous technological veneer of speech recognition and large model deployment lies a brutal covert battle centered around real-world data anchors, full-scenario lifestyle service distribution rights, and preventing oneself from being reduced to mere infrastructure.

Ride-hailing has never been just a mobility business; it is also the lowest-cost, highest-efficiency market education tool and a core pivot for lever (leveraging) the next generation of traffic entry points.

At the dawn of the mobile internet era, ride-hailing propelled the dual-oligopoly pattern (structure) of mobile payments; now, it is becoming the springboard for general-purpose AI to integrate into real life.

Part 1: Historical Polyphony—Why Ride-Hailing is the Optimal Enlightenment Entry Point

The 2014 Didi-Kuaidi battle resolved the conundrum of why users should link a social app to their bank cards.

Breaking through mental barriers cannot rely on preaching; it requires a sufficiently lightweight, high-frequency, and tangible scenario that allows users to complete their first action effortlessly.

Ride-hailing perfectly meets all these criteria: low unit prices, where a few yuan in subsidies can incentivize trial and error; instant feedback, where services are enjoyed immediately after payment, with value perceived in seconds; and high frequency, with users in first- and second-tier cities using the service at least twice weekly, rapidly forming muscle memory.

The results are already etched into industry history.

On February 7, 2014, the first workday after the Chinese New Year, Didi's peak ride orders reached 2.62 million, with WeChat Pay orders surpassing 2 million. The billions in subsidies burned by Didi, Kuaidi, Tencent, and Alibaba not only fueled the prosperity of the ride-hailing sector but also drove the widespread adoption of mobile payments in China.

Ride-hailing was merely the teaching tool; payments were the true examination.

Twelve years later, the large model industry stands at the same crossroads.

After years of iteration, parameters and reasoning capabilities of large models are no longer bottlenecks. The industry's true shortcomings are narrow user mindsets and the emptiness of practical implementation scenarios.

Most ordinary users still perceive AI as an online toy for writing copy, creating PPTs, or casual chatting, failing to generate demand for real-life applications.

No matter how much one talks about general artificial intelligence, multimodality, or Agents, it pales in comparison to a perceptible real-world experience.

Ride-hailing is the optimal solution for AI mindset enlightenment on a national scale.

Users need not learn prompt engineering, understand large model principles, or even realize they are using AI—simply speaking a sentence completes the entire operation.

The cost of trial and error is nearly zero, with prices identical to regular ride-hailing; value feedback is rigid, with address accuracy and arrival times yielding results within minutes. Combined with its high-frequency, essential nature, it rapidly embeds the cognition (cognition) that AI can get things done into users' behavioral habits.

In May 2026, nationwide ride-hailing orders reached 977 million, a transaction frequency enticing enough to captivate players.

Of course, it must be acknowledged that AI ride-hailing's offline transformations are limited—core issues like peak-hour capacity shortages, dynamic pricing fluctuations, and standardized service deficiencies remain untouched; online interaction optimizations are minimal, reducing screen taps, eliminating manual address input, and avoiding the need to switch between multiple apps for price comparisons.

Nevertheless, the significance of AI ride-hailing extends far beyond minor functional tweaks—it represents a user mindset assault replicating the historical path of mobile payments.

Ride-hailing is a litmus test, prying open not just a corner of the mobility market but also a crack in the door for general-purpose AI to enter everyday life.

Part 2: Three Players Enter—Non-Overlapping Tracks, Ulterior Motives

Unlike the head-on clash between Didi and Kuaidi back then, today's three players in AI ride-hailing appear to target the same feature but wield entirely different entry tickets and pursue vastly different destinations. This is not a price war on the same dimension but an ecological positioning battle driven by distinct needs.

Didi is the incumbent transitioning from target to player.

Twelve years ago, Didi served as the entry carrier (vehicle) for payment giants, soaring on the winds of the subsidy war to become the undisputed leader in the mobility sector. Today, for Didi, this is an unavoidable survival defense war, necessitating its direct involvement in AI.

If users grow accustomed to hailing rides via a single command in AI assistants, ceasing to proactively open the Didi app, Didi's brand equity accumulated over a decade will rapidly diminish, relegating it to a background transportation supplier—much like how banks were reduced to payment channels after mobile payment popularize (popularization).

Losing direct user touchpoints means forfeiting pricing power, brand premium, and initiative in business expansion, ultimately reducing it to a sweatshop earning meager profit margins.

Thus, Didi's core objective in AI is to safeguard user entry points, retaining interaction initiative through its densest national transportation network, most mature dispatch system, and entrenched user mindset in mobility.

For Didi, AI is an efficiency tool for its existing business and a moat against being reduced to infrastructure.

Qianwen is the orchestrator of an ecological closed loop.

Alibaba's foray into AI ride-hailing aims to invigorate its entire lifestyle services ecosystem. This represents a grand parade of consumer ecosystem AI transformation, with AutoNavi's maps and transportation, Ele.me's local dining, Fliggy's travel, Taobao's e-commerce, and Alipay's payment chain—all scattered businesses previously requiring users to switch apps—now unified.

AI ride-hailing serves as the connector for this entire ecosystem.

A user's statement, "Dine at a hotpot restaurant in XX business district at 7 PM," enables AI to simultaneously handle restaurant selection, group-buying deals, ride-hailing navigation, and even post-meal entertainment recommendations, all within a conversational flow without app-switching.

Ride-hailing, as the lowest-threshold, highest-frequency touchpoint, educates users that AI can manage entire itineraries, ultimately locking them into Alibaba's payment and service closed loop.

Doubao is the disruptor seeking a shortcut to overtake rivals.

ByteDance's entry bears the offensive spirit of WeChat Pay back in the day. Doubao's model capabilities are on par, yet it remains trapped in the content creation tool sphere. By leveraging ride-hailing, it follows WeChat Pay's playbook of using an essential scenario to pull AI from a creation tool into a lifestyle entry point, first cracking user mindsets before gradually infiltrating full-scenario services like dining, ticketing, hotels, and group-buying.

ByteDance commands vast public domain traffic from short videos and search. Once user mindsets are penetrated, traffic conversion efficiency will be staggering.

Deeper still, Alibaba and ByteDance are not betting on the 411.9 billion yuan ride-hailing market but on securing the chief entry point for AI-driven lifestyle services. This is a high-stakes gamble where ride-hailing is the ticket, and the ecological feast beyond entry is the true prize.

Three players, three paths: one defends its mobility stronghold, one orchestrates an ecological closed loop, and one competes for entry-point dominance. Superficially competing in the same arena, they each occupy distinct battlefields with their own survival stakes.

Part 3: The Entry-Point Battle—Distribution Rights Shift and the Winner-Takes-All Logic

The true entry-point competition is not about access paths but service distribution rights—this is the war's core essence.

For over a decade, the internet logic was "users seek services": users first had a clear ride-hailing need before proactively opening Didi, AutoNavi, or Baidu Maps, with platforms fulfilling orders. In this phase, vertical service platforms held the discourse power (voice), with entry points merely serving as traffic channels.

AI ride-hailing completely reverses this logic, prioritizing intention before service.

Users need not decide which platform to use or even explicitly choose ride-hailing; they simply express their final destination, with AI handling address recognition, vehicle matching, transportation selection, ordering, and payment.

This means service distribution rights shift entirely from vertical mobility platforms to AI entry points.

AI decides whether to assign orders to Didi or Cao Cao Mobility, with users perceiving no difference or caring less. Over time, traditional mobility platforms' brands will fade into obscurity, relegating them to invisible background transportation providers.

More critical than distribution rights are the dual interceptions of data and transactions.

Mobility data represents the highest-quality real-world behavioral anchors among all lifestyle services, inherently tied to time, location, spending power, and travel patterns, deeply binding to users' daily lives. With this data, AI can craft precise user lifestyle profiles, refining service matching capabilities in a virtuous cycle—a barrier online-only large models can never surpass.

Transaction chain closure means entry-point players fully control commercialization initiative.

Users complete decision-making and payment entirely within chat interfaces, eliminating the need to switch to service platforms. The lion's share of commissions, advertising, and value-added service revenues will ultimately concentrate at the entry point.

Notably, AI entry points inherently possess strong exclusivity. Just as users retain only one primary input method or browser, they will fixate on one AI assistant for daily needs. Once path dependency forms—where users instinctively turn to it for all lifestyle demands—all such needs will converge there, culminating in a winner-takes-all scenario.

This also dictates that the war's outcome hinges not on the mobility sector itself.

Winning the ride-hailing scene merely grants entry; the true decisive battle lies in extending ride-hailing habits to broader lifestyle services like dining, travel, and retail.

Yet, this sector harbors a massive hidden concern. Mobility is a heavily offline, regulated, and service-intensive domain, with a vastly different tolerance for error compared to pure online text tools.

Most large model firms lack offline service operational DNA.

Qianwen currently relies on AutoNavi's aggregated transportation as a safety net, while Doubao outsources fulfillment to Cao Cao Mobility—both effectively outsourcing offline services. Didi alone shoulders all costs, from model training to driver training, safety management, and complaint handling.

In the short term, the aggregation model appears nimble and appealing; however, as service consistency, safety baselines, and crisis response become critical differentiators long-term, Didi—with its proprietary transportation and direct management experience—will likely widen its substantive lead over outsourcing-dependent rivals.

This hidden fault line will gradually emerge as AI ride-hailing transitions from novelty to routine use.

For now, no absolute winner will emerge from this AI ride-hailing race.

Long-term differentiation will arise along two dimensions: who can transform AI ride-hailing from novelty into habit, forming stable path dependency, and who can perfectly balance AI's lightweight online decision-making with the heavyweight offline fulfillment.

AI ride-hailing is no mobility revolution. At its core, it is merely another battleground in the internet's relentless traffic wars—shifting from text to short videos, short videos to live-stream commerce, live-stream commerce to instant retail, and now to instant mobility.

Twelve years ago, a ride-hailing war delivered mobile payments to billions' phones, inaugurating China's mobile internet golden decade. Twelve years later, the same script repeats, but this time, the objective is to popularize AI-driven lifestyle orchestration.

A simple ride-hailing command is a door.

On one side lies the app-cluttered mobile internet era; on the other, a conversational AI-led next-generation lifestyle services ecosystem.

Technology dons a new costume every few years, but the stories of contend for (vying) for entry points, data, and user decision-making rights remain unchanged.

END

Wang Qingru @ okokok-74

Long-term observer of internet giants and vertical industry leaders. Welcome to connect and communicate.