Recently, there has been plenty of good news for Apple.

Just before, Apple announced that Apple Intelligence is about to launch in China, integrating Alibaba's Tongyi Qianwen large model, opening testing on November 4 and officially landing with iOS 27.3 in December. Shortly after, Apple delivered its fourth fiscal quarter results for fiscal 2026: revenue and earnings per share both exceeded market expectations, and iPhone revenue hit a record high for the same period in history, drawing a perfect conclusion to the "best fiscal year ever."

However, new CEO John Ternus is not satisfied with Apple's current situation. According to reports, last week, in a closed-door meeting room at Apple Park, he issued an unprecedentedly stern warning to the software team:

Apple has fallen far behind competitors in AI and must embrace AI with the same urgency as its rivals, accelerate the roadmap, and quickly move beyond the stage of "merely making necessary patches to Siri."

This is probably Apple's real good news—the new leader is not content with winning by doing nothing.

Over the past year, Apple's inaction in AI has outperformed most companies; even the storage storm triggered by AI further deepened iPhone's share in the high-end market. But AI is fundamentally changing the value of smartphones, user entry points, and the profit models of tech companies, and this is precisely Apple's core business value.

As the definer of smartphones, Apple is the biggest beneficiary of the mobile internet era. But at this very moment, it seems to be in the same situation as Nokia was back then.

01 Anxiety in the earnings report

In the fourth fiscal quarter of fiscal 2026 (the third quarter of the calendar year), Apple achieved revenue of $114.5 billion, up 12% year over year, also Apple's highest-revenue fourth fiscal quarter in history; net profit was $30.7 billion, up 12% year over year.

In addition, iPhone revenue in this fiscal quarter was $57.58 billion, up 17.5% year over year, setting a record high for the same period in history; services revenue was $31.57 billion, up 9.8% year over year, a record high. New CEO Ternus said, "This is the best-performing fiscal year we have ever had." Especially at present, when Wall Street tech giants are struggling with AI capital expenditures, Apple's report card looks even more impressive.

But when examined closely, Apple's problems are also obvious.

Over the past year, the iPhone 17 series' product strategy of "more for the same price" allowed it to capture more upgrade demand amid the global wave of phone price increases; at the same time, the iPhone series began raising prices in September, which also drove further growth in iPhone revenue. But as the storage price increase cycle extends further, Apple can hardly remain unaffected.

On the one hand, profits are being gradually eroded. Research firm TrendForce estimates that using the 256GB iPhone 18 Pro as a sample, the overall BOM material cost is about 38% higher than that of the Pro model in the same period of 2025. Management said it expects gross margin (45%-46%) in the next fiscal quarter to be lower than in this fiscal quarter.

On the other hand, sales of other hardware have begun to decline significantly. According to Omdia data, global iPad shipments in the third quarter of 2026 were about 11.5 million to 11.58 million units, plunging 27.5% year over year, a decline larger than the global tablet market's average drop of 23%. Data from RUNTO shows that in the third quarter of 2026, iPad sales on Chinese e-commerce platforms fell 45.2% year over year, and market share decreased by 8.5 percentage points year over year to 19.1%.

At the same time, growth in Apple's most important profit source in the past—services revenue—has begun to slow. Although Apple's services revenue reached $31.57 billion in this fiscal quarter, a record high, its growth rate was only 9.8%, the first time in nearly three years that growth fell below 10%.

Management said the slowdown in services revenue growth was mainly due to exchange rates; but in reality, under regional regulatory requirements, Apple's commission rates and payment channel restrictions have loosened in many places. For example, starting October 1, the EU App Store commission will drop from 30% to 26%, with lower rates applying to some eligible developers.

But these are only appetizers. The biggest challenge is still AI.

02 "Intelligence" arrives late

In June 2024, Apple Intelligence made a stunning debut at WWDC. But by the time Chinese users can use it, two years and four months will have passed.

Disappointingly, although it is a new feature arriving late after two years, Apple's time seems to have stood still two years ago.

In the layout of on-device AI, domestic manufacturers have already made many innovative attempts. For example, Xiaomi is trying to use phones to connect an intelligent ecosystem, while Nubia's Doubao phone reshapes interaction logic through GUI. AI assistants are all trying long-chain operations to solve user needs in one step.

But the first batch of intelligent features Apple has served up mainly includes photo erasure, real-time translation, Genmoji, and writing proofreading. They do not yet have multi-task orchestration and cross-app execution capabilities, and remain an old solution of "single request, single feedback."

It is not that Apple does not understand the importance of AI.

Over the past decade or more, the smartphone business model has had a very important characteristic: whoever controls the operating system controls the entry point for users into the digital world. This is also the fundamental reason Apple can control app distribution channels and earn revenue through the App Store, payments, and subscription businesses.

And AI is changing this structure. On-device AI is gradually evolving into a window that can solve problems in one stop. Previously, users needed to know which app to go to for a service; now users may only need to know which AI assistant to make a request to. This leads to the migration of entry points.

But at present, there are two main reasons troubling Apple's AI progress.

One is that Apple's progress in research and development speed, model capabilities, and technical talent for cutting-edge foundation models completely fails to match its position in hardware and operating systems. This has forced Apple to rely on external technology to advance AI—first OpenAI, then Google, and domestically Alibaba and Baidu.

The second is Apple's insistence on privacy protection.

The demonstration of AI capabilities on the device itself requires users to give up part of their personal privacy, using long-term memory to help you plan and make decisions.

But for so many years, Apple has always left all decision-making power to users, so much so that even for operations like accessing the clipboard, Apple keeps popping up windows to obtain your consent. This inertia has made Apple's software chief Federighi long maintain a cautious attitude toward advancing generative AI.

According to Bloomberg, Federighi has an extremely strong style within the software team and is accustomed to using the technical details he controls to push his positions in meetings, making it difficult for many more radical explorations within the team to proceed.

Lagging behind in foundation models has also intensified this conflict. Apple needs models from companies such as Google and OpenAI to make up for capabilities, but because these involve user data security and core technological autonomy, it cannot fully hand over the product. During WWDC 2024, Federighi said in an interview with the media when talking about the integration of ChatGPT and Siri, "Before switching to ChatGPT, we will ask for your consent first," while also saying, "From a privacy perspective, users always have control, and the entire process is transparent."

03 The "new direction" in the AI era

Apple has already been adjusting the pace of AI.

In addition to the closed-door meeting mentioned at the beginning, Bloomberg disclosed that in the development of the next-generation system codenamed iOS 28, Apple will officially adjust the first phase (M1), which has continued the years-long model of "4 weeks coding + 4 weeks system integration," to "6 weeks coding exploration + 2 weeks system integration." Moving the focus clearly earlier is precisely to give underlying engineers time for free experimentation and force the team to explore radical Agent architectures.

Another focus is the hardware ecosystem.

In an analysis published on April 22, 2026, Reuters pointed out that Apple achieved great success in the past through a highly controlled ecosystem, but AI competitors relying on open models and developer ecosystems to innovate quickly may change the rules of competition. Generative AI has changed the way users interact with the digital world. In the future, users may only need to make requests to an AI assistant, and the Agent will automatically call different applications to complete tasks.

Therefore, Apple needs to extend the iPhone's ecosystem advantages into the home space, enhance user stickiness, and thereby control more entry points.

In the second half of the year, Apple began participating in the definition and design of third-party smart home products, partnering with brands such as LG Electronics and Schneider Electric to launch the Apple Home Collection. Products include smart video doorbells, smart door locks, indoor cameras, room sensors, and more.

Of course, Apple still has sufficiently strong advantages in hardware, chips, and ecosystem, but these advantages cannot determine Apple's fate in the AI era.

In addition to catching up in foundation model capabilities, Apple more needs to break out of the R&D and product inertia established over the past decade or more and enter the rhythm that the new era demands.

References:

1. Apple earnings report

2. ifanr: Ternus pressures Siri chief, Apple AI cannot wait any longer

3. Reuters: In the AI era, Apple's strengths may become its constraints