文 | 文武赵
10月26日，苹果发布iOS 27.2系统更新，继续扩大Siri AI的能力。
这次更新中，Siri增加了对历史对话的记忆功能，能够调用语音邮件、信息附件等更多个人数据。与此同时，苹果还增加了“排除应用”选项，允许用户决定哪些App可以被Siri访问。
显然，苹果正在让Siri越来越深入地介入用户的数字生活。从读取邮件、理解屏幕内容，到调用应用完成任务，Siri正逐渐获得过去由一个个独立App承担的功能。
一、苹果最赚钱的生意，建立在App之上
苹果是全球移动互联网商业秩序的重要建立者。
2008年，第一代App Store上线。开发者获得了触达全球iPhone用户的渠道，消费者则可以通过统一的商店寻找、下载和购买应用。
此后，App成为移动互联网最重要的商业单元。
购物需要打开亚马逊，打车需要Uber，订酒店需要Booking，听音乐需要Spotify。
在中国市场，类似的商业结构体现在淘宝、美团、携程、抖音和微信等超级应用中。
每一个App都试图争夺用户的时间、交易和支付行为。
苹果则处在这些商业活动的上游。
2026年6月，苹果公布的一项研究显示，2025年全球App Store生态系统促成的开发者账单金额和销售额超过1.4万亿美元。
其中，实体商品和服务销售额达到1.1万亿美元，数字商品和服务约1490亿美元，应用内广告收入约1510亿美元。
需要说明的是，这1.4万亿美元属于整个App生态的商业规模，苹果并不能从所有交易中抽取佣金。苹果披露，超过90%的相关交易金额没有向其支付佣金。
但这一数据仍然说明，苹果建立了一套覆盖商品交易、数字消费和广告业务的庞大商业基础设施。
这套体系还在持续增长。
2026财年前三个季度，苹果服务业务收入达到917.28亿美元，同比增长14%。
按照第三季度财务数据计算，服务业务贡献了约232.45亿美元毛利润，占公司总毛利润的42.4%。
虽然苹果并未单独披露App Store的完整收入和利润，但应用分发、广告、订阅、云服务等业务共同支撑了这一高利润结构。
它们拥有一个共同的商业基础：iPhone持续充当用户访问数字服务的重要入口。
苹果通过操作系统和应用商店组织这些服务，开发者则在各自的应用内部经营生意。
谁控制应用分发，谁就能够影响移动互联网的流量和交易规则。
然而，AI正在尝试改变这种关系。
二、Siri正在让用户不必打开App
2026年9月14日，苹果正式推出新版Siri AI。
这是Apple Intelligence发布两年后，苹果在人工智能领域一次重要的产品升级。
新版Siri具备个人信息理解、屏幕内容感知、跨应用操作等能力。苹果还推出了独立的Siri应用，支持保存历史对话，并在不同苹果设备之间继续交流。
其底层模型采用了苹果与谷歌Gemini合作开发的新一代Apple Foundation Models。
从产品功能看，Siri正在逐渐从语音助手变成任务执行入口。
假设一名用户准备周末去纽约旅行。
过去，他可能先在航旅平台查询机票，再通过Booking预订酒店，打开地图规划行程，最后使用餐饮平台寻找餐厅。
整个过程需要在多个App之间切换。
如果Siri能够完成这些操作，用户只需要提出一个需求：
“帮我安排一个周末纽约旅行，预算1500美元，酒店靠近中央公园。”
AI可以根据用户偏好搜索信息，比较服务，并在获得授权后调用相关应用。
上述完整旅行预订仍属于AI Agent的发展方向，不能视为当前Siri已经普遍实现的功能。但苹果正在为这种交互方式建立基础。
其中一项关键技术是App Intents。
开发者可以通过App Intents，把应用中的功能以结构化方式提供给Siri、Spotlight和快捷指令等系统服务。
一款购物应用可以开放商品查询，一款音乐应用可以开放播放功能，一款任务管理应用可以开放创建待办事项的能力。
随着开发者进一步适配，App可以从用户主动打开的软件，变成AI按需调用的后台服务。
这一变化看似只是减少了操作步骤，商业影响却可能更加深远。
过去，用户需要决定打开哪个App。
未来，AI可能替用户作出选择。
例如，用户要求预订酒店，究竟由Booking、Expedia还是其他平台完成？
用户准备购买耳机，AI会优先调用亚马逊、品牌官方网站，还是其他零售渠道？
当AI拥有了选择服务商的能力，它就开始参与移动互联网的流量分配。
过去，苹果分发App；未来，Siri可能直接分发订单。
这可能是苹果AI战略中商业价值最高的一步。
它也意味着，苹果建立的传统应用入口正在被自己的产品改变。
三、Siri越强大，开发者可能越不赚钱
这一变化最先影响的，可能是依靠用户流量和广告收入生存的开发者。
传统App拥有一套完整的商业化流程。
用户打开应用，浏览页面，看到广告和推荐内容，完成购买或者订阅。平台可以记录用户行为，并通过推荐算法、会员体系和广告产品提高收入。
对于电商平台来说，商品搜索和推荐排序是重要的广告资源。
对于在线旅游平台来说，酒店排序、机票推荐以及附加服务构成了收入来源。
对于内容平台来说，用户停留时长决定了广告展示和商业化空间。
但当Siri代替用户完成越来越多任务，这些应用可能失去一部分直接接触消费者的机会。
用户只需要获得最终结果。
谁提供服务、平台展示了什么广告、还有哪些商品被推荐，可能都不再重要。
假设一家酒店预订平台原本通过搜索广告、推荐位置和会员服务获得收入。
如果用户直接通过Siri完成酒店比较和预订，平台仍然可能获得订单，但它在交易过程中的广告展示机会可能减少。
当然，AI带来的订单也可能降低平台的获客成本，甚至增加交易转化率。
问题在于，新增交易收益能否弥补原有流量价值的损失。
苹果面临的挑战也因此出现。
Siri要变得有用，就需要更多第三方App开放功能和数据。
但开发者可能担心，开放得越充分，自己对用户的控制力越弱。
更进一步，AI可能改变开发者之间的竞争方式。
过去，开发者争夺App Store搜索排名、下载量和用户使用时长。
未来，他们还可能争夺Siri的推荐和调用机会。
一旦用户把选择权交给AI，决定调用哪家服务商的规则就会具有商业价值。
这些规则可能涉及价格、用户偏好、服务质量、隐私授权以及商业合作。
苹果目前尚未公布成熟的AI任务分发商业体系。
但这种机制如果大规模建立，开发者和苹果之间的利益分配就可能发生变化。
值得注意的是，苹果自身也面临类似的商业压力。
2026财年第三季度，苹果披露服务业务增长主要来自广告和云服务。
如果AI逐渐减少传统应用内的浏览、搜索与广告展示，部分广告商业模式可能受到冲击。
不过，苹果服务收入还包括iCloud、Apple Music等业务，App Store佣金也并不覆盖所有交易。AI是否最终压缩苹果自身的服务利润，目前仍缺少直接证据。
更合理的判断是，苹果正在尝试把一部分商业价值，从传统应用分发环节迁移到AI任务分配环节。
过去开发者争夺的是用户点击，未来可能争夺AI调用。
谁掌握任务入口，谁就有机会建立新的商业规则。
四、苹果为什么必须亲手拆掉旧生意？
苹果的选择背后，还有一个更加现实的竞争压力。
它并非唯一能够提供AI助手的公司。
OpenAI拥有ChatGPT，谷歌拥有Gemini，其他科技公司也在推动具备工具调用和任务执行能力的AI产品。
这些产品可以直接运行在iPhone上。
如果未来用户习惯打开ChatGPT或者Gemini完成旅行规划、购物、搜索和办公任务，即使这些交易最终仍然发生在iPhone上，苹果也可能逐渐失去对用户需求的直接控制。
这将带来一种新的产业分工。
苹果提供硬件和操作系统，第三方AI助手掌握用户意图，再决定调用哪些应用和服务。
苹果可能仍然销售大量iPhone，却无法像过去一样决定用户从哪里开始数字消费。
这种风险解释了苹果为何急于升级Siri。
2026年1月，苹果与谷歌达成多年期合作，以Gemini技术作为下一代Apple Foundation Models的重要基础。
这意味着苹果可以通过外部合作获得先进模型能力，把更多资源投入操作系统、个人信息管理和应用调用能力。
苹果长期擅长这种商业分工。
在搜索时代，它没有自行建立一个与谷歌规模相当的全球搜索引擎，却通过Safari默认搜索合作获得巨额收入。
但AI时代有一个重要区别。
搜索引擎通常负责提供信息和链接，用户仍然需要进入网站或App完成后续操作。
AI Agent则有机会直接参与交易和任务执行。
如果苹果能够让Siri成为默认任务入口，模型来自谷歌还是苹果自身，对它的商业控制力未必具有决定性影响。
如果用户习惯通过其他AI助手发出指令，那么苹果即便拥有操作系统和App Store，也可能无法掌握最关键的需求分配权。
截至2026年10月，新版Siri仍处于逐步推广阶段，中国大陆尚未开放，部分跨应用能力还受到开发者适配、隐私权限和系统执行能力的限制。
苹果也没有公布足够的数据，证明Siri已经能够稳定完成大规模复杂交易。
因此，目前谈论App Store被AI取代仍然过早。
但苹果已经开始推动一场可能改变自身商业模式的转型。
2008年，App Store让苹果成为移动互联网应用分发的中心。
十八年后，Siri试图让用户绕过应用图标，直接通过AI完成需求。
在这个过程中，App不会很快消失，但它可能逐渐退居后台。
苹果最需要守住的，是用户发出指令之后，谁来决定调用什么服务。
从这个角度看，苹果正在主动削弱自己建立的旧入口，以争夺一个更有价值的新入口。
这也构成了Siri AI最大的商业悖论。
如果Siri成功，苹果可能需要重新调整过去依靠App分发、应用内交易和开发者生态建立的部分收益结构。
如果Siri失败，用户需求可能越来越多地被第三方AI助手接管。
苹果已经没有太多维持原状的空间。
对于一家每季度服务业务收入超过300亿美元、毛利率达到75.6%的公司来说，主动改变现有商业秩序需要付出代价。
但苹果更担心的，可能是未来仍然拥有iPhone和App Store，却失去了决定用户打开什么、购买什么以及使用什么服务的权力。
苹果亲手拆掉的，可能是App时代的流量入口；它希望得到的，则是AI时代的任务分配权。
本文由 AI宝库自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
Wen Wen Wu Zhao
On 26 October, Apple publishiOS 27.2 was updated to continue to expand the capacity of Siri AI.
In this update, Siri has added memory features to the historical dialogue and has been able to access additional personal data, such as voice mail, information attachments, etc. At the same time, apples have added the “exclusion application” option, allowing users to decide which Apps can be accessed by Siri.
It is clear that apples are increasingly involving Siri in user digital life. From reading mail, understanding screen content, to calling applications to do their job, Siri is gradually acquiring functions previously performed by an independent App.
One: Apple's most profitable business, based on App.
Apple is an important builder of the global mobile Internet business order.
In 2008, the first generation of App Store went online. Developers have access to global iPhone users and consumers can find, download and purchase applications through a single shop.
Since then, App has become the most important business module for mobile Internet.
Shopping requires the opening of the Amazon, uber for taxis, Booking for hotels and Spotify for music.
In the Chinese market, similar business structures are found in super-applications such as poaching, beauty missions, transport, Douyin and WeChat.
Each App seeks to compete for the time, trade and payment of the user.
Apples are upstream of these commercial activities.
In June 2026, a study published by Apple showed that the global developer billing and sales of App Store ecosystems exceeded $1.4 trillion in 2025.
Of this amount, real sales of goods and services amounted to $1.1 trillion, digital goods and services to about $149 billion, and advertising revenues from applications to about $15.1 billion.
It should be noted that the $1.4 trillion is on a commercial scale for the entire App ecosystem and that apples do not draw commissions from all transactions. Apple disclosed that no commission was paid to more than 90 per cent of the relevant transaction amounts.
However, this data still indicates that apples have a large business infrastructure covering commodity transactions, digital consumption and advertising.
The system continues to grow.
In the three quarters of FY 2026, revenue from apple services reached $91.728 billion, an increase of 14 per cent over the previous year.
According to financial data for the third quarter, service operations contributed some $23,245 million, or 42.4 per cent of the company ' s gross profits.
While Apple did not disclose the full revenues and profits of App Store separately, this high-profit structure was underpinned by such operations as application distribution, advertising, subscriptions, cloud services, etc.
They have a common business base: iPhone continues to serve as an important entry point for users to access digital services.
Apples organize these services through operating systems and application shops, while developers operate their own in-house businesses.
Whoever controls the distribution of applications can influence mobile Internet traffic and transaction rules.
However, AI is trying to change this relationship.
ii. Siri is making users not need to open App
On September 14, 2026, Apple officially launched the new version of Siri AI.
Two years after Apple Intelligencepublish, Apple is a major product upgrade in artificial intelligence.
The new version of Siri has skills in personal information understanding, screen content awareness, cross-application. Apples also introduced independent Siri applications to support the preservation of historical dialogue and continued communication between different apple devices.
The bottom model uses a new generation of Apple Foundation Models, developed in collaboration with Google Gemini.
In terms of product functionality, Siri is gradually moving from voice assistants to mission implementation portals.
A user is supposed to travel to New York on weekends.
In the past, he may have searched for tickets on the flight platform, then booked hotels through Booking, opened the map to plan the trip and finally used the catering platform to find a restaurant.
The whole process needs to switch between multiple Apps.
If Siri can complete these operations, the user needs only one request:
“Set me a weekend trip to New York, with a budget of $1,500, near Central Park.”
AI can search for information based on user preferences, compare services and call relevant applications when authorized.
The complete travel booking described above remains in the direction of AI Agent and cannot be considered to be a function that is now widely realized in Siri. But apples are building the foundation for this interaction.
One of the key technologies is App Intens.
The developers can use App Intens to structure the functionality in application to provide system services such as Siri, Spotlight and fast-track commands.
A shopping application can open up commodity queries, a music application can open up play functions and a task management application can open up the ability to create to-do matters.
As the developers become more appropriate, App can turn the software from the user-initiated to AI-based back-office service.
While this change appears to have reduced operational steps, the commercial impact may be far-reaching.
In the past, users need to decide which App to open.
In the future, AI may choose between users.
For example, does the user request hotel reservations, done by Booking, Expedia or other platforms?
Are users ready to buy headphones, and AI will give priority to Amazon, brand official websites, or other retail channels?
When AI has the capacity to select service providers, it begins to participate in the distribution of mobile Internet traffic.
In the past, apples distributed App; in the future, Siri might distribute orders directly.
This is probably the highest commercial value in the Apple AI strategy.
It also means that the traditional applications of apples are being transformed by their own products.
iii. The stronger Siri, the less profitable the developers may be.
This change may be the first to affect developers who depend on user flows and advertising revenues.
Traditional App has a complete commercialization process.
Users open applications, view pages, see advertisements and recommendations, complete purchases or subscriptions. The platform can document user behaviour and increase income through recommended algorithms, membership systems and advertising products.
Commodity search and recommendation sequencing are important advertising resources for electrical platforms.
For online tourism platforms, hotel sequencing, airline tickets and additional services constitute a source of income.
For content platforms, the time spent by users determines the space for advertising and commercialization.
But when Siri replaces users with an increasing number of tasks, these applications may lose some of their direct contact with consumers.
The user needs only the final result.
It may no longer matter who provides the services, what the platform displays and what other commodities are recommended.
It is assumed that a hotel reservation platform would have earned income by searching for advertisements, recommended locations and membership services.
If a user completes a hotel comparison and booking directly through Siri, the platform may still obtain an order, but its advertising opportunities during the course of the transaction may be reduced.
Of course, the orders that AI brings can also reduce the cost of getting visitors to the platform and even increase the rate of transformation.
The question is whether the proceeds of the new transactions can compensate for the loss of the value of the original flow.
The challenges of apples also arise.
To be useful, Siri needs more third-party App open functionality and data.
However, developers may fear that the more open they become, the weaker their own control over users.
Further, AI may change the way the developers compete.
In the past, developers competed for App Store search rankings, downloads and user time.
In the future, they may also compete for Siri ' s referral and call opportunities.
The rules that determine which service provider to call will have a commercial value once the user hands over the choice to AI.
These rules may relate to price, user preferences, quality of service, privacy authorization and business cooperation.
Apple has not yet published a mature AI task distribution business system.
However, if such mechanisms are established on a large scale, the distribution of benefits between developers and apples may change.
Notably, apples themselves face similar commercial pressures.
In the third quarter of fiscal year 2026, the growth of Apple Disclosure Services was mainly due to advertising and cloud services.
Some commercial models of advertising may be affected if AI gradually reduces browsing, searching and advertising in traditional applications.
However, Apple services revenue also includes iClaud, Apple Music, and App Store commissions do not cover all transactions. Whether AI ultimately compresses the service profits of Apple itself still lacks direct evidence.
More rationally, apples are trying to move some of their commercial values from traditional applications distribution to AI tasking.
In the past, developers competed for user clicks, and in the future they might compete for calls from AI.
Whoever controls the mission entrance has the opportunity to establish new business rules.
Four: Why must apples tear down old business?
There is a more realistic competitive pressure behind the choice of apples.
It is not the only company that can provide AI assistants.
Open AI has ChatGPT, Google has Gemini, and other technology companies are promoting AI products with tool call and mission performance capabilities.
These products can run directly on iPhone.
If future users become accustomed to opening ChatGPT or Gemini to travel planning, shopping, search and office tasks, apples may gradually lose direct control of user needs even if these transactions eventually continue on iPhone.
This will lead to a new industrial division of labour.
Apple provides hardware and operating systems, and third-party AI assistants have user intentions before deciding which applications and services to call.
Apples may still sell large amounts of iPhone but cannot determine where users start digital consumption as in the past.
This risk explains why apples are rushing to upgrade Siri.
In January 2026, Apple entered into multi-year collaboration with Google to use Gemini technology as an important basis for the next generation of Apple Foundation Models.
This means that apples can obtain advanced modelling capabilities through external cooperation, and can invest more resources in operating systems, personal information management and application mobilization capabilities.
Apple has long been good at this commercial division of labour.
In the search age, it had not built its own global search engine on a scale comparable to Google, but it had earned huge revenues through Safari ' s default search cooperation.
But there is an important difference between the AI era.
The search engine is usually responsible for providing information and links, and users still need to access the website or the App for follow-up.
AI Agent has the opportunity to be directly involved in transactions and mission execution.
If apples can make Siri the default mission portal, the model comes from Google or Apple itself, and it may not have a decisive impact on its commercial control.
If users are used to issuing instructions through other AI assistants, apples may not have access to the most critical distribution of needs even if they have an operating system and App Store.
As of October 2026, the new version of Siri was in the phase of gradual expansion, the mainland of China was not yet open, and some of the cross-application capabilities were limited by the ability of developers to adapt, privacy permission and system implementation.
Apples have also not published sufficient data to prove that Siri has been able to stabilize large-scale and complex transactions.
So it is still too early to talk about App Store being replaced by AI.
But apples have begun to drive a transformation that could change their own business model.
In 2008, App Store made apples the centre for distribution of mobile Internet applications.
Eighteen years later, Siri tried to get users to bypass the application icon and complete the demand directly through AI.
In the process, App will not disappear soon, but it may step back.
The most important thing for apples is to decide which service to call after the user has given the order.
From this perspective, apples are proactively weakening their own old entrances in order to compete for a new and more valuable entrance.
It also constitutes the biggest commercial paradox of Siri AI.
If Siri succeeds, apples may need to realign some of the revenue structures that were built in the past on App distribution, application trading and developers ' ecology.
If Siri fails, user needs may increasingly be taken over by third-party AI assistants.
Apples don't have much space to maintain their status quo ante.
For a company with a quarterly service income of over $30 billion and a Maori ratio of 75.6 per cent, it would be costly to proactively change the existing business order.
But apples are more worried that they may still have iPhone and App Store in the future, but lose the power to decide what users open, what to buy and what to use.
Apple itself is probably the entry point of the App era; what it wants is for the AI era.
Auto-fetched and translated by AI Gem Box. All rights belong to the original author.