During intraday trading on October 30, Apple's market cap hit $5 trillion, knocking NVIDIA off its throne.

By the close the next day, the dust had settled: Apple at $5.03 trillion, NVIDIA at $4.93 trillion. NVIDIA shed $217 billion in a single day—roughly RMB 1.4 trillion, more than PetroChina's entire market value wiped out in one session.

Image source: Bloomberg

This is not the first time Apple has surpassed NVIDIA. As early as July 17 this year, Apple had already overtaken NVIDIA intraday, and on July 27 it led again—though it failed to hold the top spot both times. Since mid-2025, NVIDIA had dominated this ranking for more than 260 trading days.

So how long this summit will last remains unknown.

But what is truly worth pondering is another matter: the company that took the crown as the biggest AI winner is the one that least resembles an AI player.

Apple is up 27% this year, the top gainer among the Magnificent Seven and more than double second-place NVIDIA. A company that has been criticized for two years as 'behind in AI' outperformed all AI companies in AI's hottest year.

Image source: Internet

This is illogical. Unless our understanding of 'how AI makes money' has been wrong from the very beginning.

It's not that Apple got stronger—it's that AI got too expensive

Let's be clear: in this round of throne-changing, Apple did not get stronger—AI got more expensive.

The market narrative is shifting from 'who has the largest investment numbers' to 'who can turn investment into profit.'

One set of comparisons is enough. HSBC estimates that AI data center capital expenditure will account for about 2.5% of Apple's projected 2026 sales. For Meta, Google, Amazon, and Microsoft, that figure is 39%.

Microsoft, Google, Amazon, and Meta each have capital expenditure at the hundred-billion-dollar level. Apple's AI-related spending in 2026 is about $11 billion to $13 billion—less than a fraction of theirs.

Image source: Internet

Spending the money is easy, but paying it back is not. Free cash flow at Alphabet and Amazon has turned negative, forcing them to increasingly turn to bond market financing. Interest rates are still rising, and borrowing costs are climbing with them. S&P directly downgraded Meta's outlook to negative, citing 'significant uncertainty over capital expenditure and free operating cash flow.'

Of course, the most intuitive example is the 'collective discount on U.S. tech stocks' on the last trading day of October. Intel fell 28.58% in a single day, its largest one-day drop in more than four decades; the Philadelphia Semiconductor Index fell 7.2%, its largest drop since June; CoreWeave and Nebius, which provide computing power leasing for AI, both fell more than 30% in October.

The Nasdaq 100 fell 6.7% in October, its worst month since March 2025.

On Apple's side, free cash flow is expected to grow 38% to $142 billion in fiscal 2026. It holds about $146 billion in cash and securities; the stock is up 27% year to date, first among the Magnificent Seven.

Image source: Internet

On one side, burning cash to the point of issuing debt; on the other, cash piled into mountains. Capital is flowing from the computing power narrative to the cash flow narrative—a shift from favoring stories to returning to basic business logic.

Without heavy AI bets, why is it being repriced?

The premium the market gives Apple mainly comes from three layers of logic.

The first layer: safe-haven attributes. Among the Magnificent Seven, Apple is the only 'quasi-utility' company. Its revenue does not depend on who wins a certain technology cycle; it depends only on 2.5 billion devices still being in users' hands. When everyone is betting on AI's return cycle, predictability itself becomes scarce. Tony Meadows, head of investment at BRI in the UK, put it bluntly: Apple has lower exposure to capital-expenditure-intensive models, and this repricing reflects confidence in earnings sustainability rather than speculation on AI prospects.

Image source: Internet

The second layer: indirect monetization. Apple is a rare 'low-input beneficiary.' No matter whether OpenAI, Google, or Alibaba ultimately wins, AI must pass through Apple's gateway to reach users. It does not compete, but it is the track itself.

The third layer: financial structure. Services gross margin is 75.6%, products gross margin about 40%. Hardware brings people in the door; services collect the money. This structure makes Apple's AI investment naturally 'light': on-device models run on chips users have already bought, and computing costs have long been paid for by hardware.

But don't rush to be bullish—Apple has many problems of its own.

First, valuation is not cheap. TTM P/E is in the 30x to 36x range, close to 36x based on expected earnings over the next 12 months, the highest level since 2008 and at the 99th percentile of its historical valuation distribution. It has become the only stock among the Magnificent Seven with a negative implied return over the next 12 months. Since the earnings report at the end of July, at least six analysts have downgraded it, and the proportion of buy recommendations is below 60%, while Microsoft, Amazon, and Meta all exceed 90%.

Siri is no longer delayed, but rolling it out globally is not easy. The new Siri AI launched in Beta with iOS 27 on September 14, 2026, supporting 16 languages, including Simplified Chinese, but mainland Chinese users did not get access to some AI feature Beta versions until November 4; however, the full experience with true cross-app autonomous invocation capabilities will not arrive until next month.

Apple's hyped 'Visual Intelligence' is powered by Google Gemini behind the scenes. Apple pays about $1 billion annually to license a 1.2-trillion-parameter model, with complex computing hosted on Google Cloud, while the Chinese version switches to Alibaba's Qwen.

Image source: Internet

Some AI capabilities are essentially 'outsourced.' In January 2026, Google and Apple officially announced a multi-year strategic partnership, with the next generation of Apple foundation models built directly on Gemini's model architecture and cloud infrastructure, while Apple's self-developed trillion-parameter model is still on the way.

So the core judgment is just one sentence: the market is buying Apple not because it does AI best, but because it does not have to burn money on AI. These two things have completely different valuation implications.

Can Apple carve out a money-making path independent of AI?

Yes, and it has already done so three times.

This playbook has a name: the 'toll booth model'—don't build the road, just set up the checkpoint.

During the cloud computing boom, Apple did not build its own; it bought Amazon AWS and resold it to users, earning the spread. In the search era, it did not develop its own; it signed a pre-installation agreement with Google and collected tens of billions of dollars a year. Now that AI has arrived, it outsources models to OpenAI, Google, and Alibaba while guarding the gateway itself.

Three times, not a single miss.

Institutional forecasts are not bad either: FY2028 revenue of $534.1 billion, services CAGR of 12.4%, the fastest-growing and highest-quality profit segment in the revenue structure.

But the ceiling is equally clear.

First, this model is strongly tied to the iPhone upgrade cycle. If phones don't sell, no cars pass through the toll booth. Second, antitrust pressure on App Store commissions is ongoing, and regulators in Europe and the U.S. have no intention of letting go.

Image source: Internet

So this path can work, but it is 'steady-state rent collection,' not a 'high-growth story.'

What does that mean? It means that viewing Apple through a tech-stock valuation framework will always make it seem 'not sexy enough'; viewing it through a consumer-stock framework, it is instead one of the highest-quality cash flow assets in the world. This round of market repricing is essentially a switch in valuation coordinates.

Global market strong, Chinese market weak

Apple tops the world, yet in China a set of the most contradictory readings has emerged: revenue growth ranks first globally, while China market share is relatively weak.

In Q3 this year, Apple shipped 11.6 million units in China, up 15% year on year, for a 20% share. Huawei reached 12.4 million units in Q3, up 18%, with share rising to 21%. Together the two accounted for 41% of the market, about 10 percentage points higher than a year earlier. The market is concentrating toward a duopoly, and Apple has gone from sole king to one of two strong players.

Image source: Internet

According to authoritative IDC data, iPhone shipments in the quarter were about 10.33 million units, down 4.3% year on year, yet share rose from 15.8% to 17.8%. Volume is falling while share is rising, indicating the overall pie is shrinking even faster.

Moreover, Apple's price bands are beginning to fall. In the mainstream high-end segment of RMB 5,000 to 6,000, Huawei has overtaken Apple to take first place; above RMB 8,000, Huawei's share exceeded 30% for the first time, a record high. Apple still holds the overall market above RMB 6,000 and an absolute 73.5% advantage above RMB 8,000, but it is losing the mid-tier.

Revenue is rising, share is being conceded, and Apple can only use price to trade for volume.

During 618, the iPhone 17 Pro series was cut by RMB 1,000 outright, with subsidies saving up to RMB 2,000, and the standard version fell below RMB 4,500 for the first time. Huawei responded the same day, cutting Mate X6 by RMB 3,000. By September this year, iPhone 18 Pro/Pro Max each rose by $100, the base model was postponed to next spring, and only two high-end models were on sale in autumn. Volume could not rise, but average selling price did.

Apple in China is not unable to sell, but selling expensive and defending with fatigue. In the past it stood on brand premium to sell; now it must rely on price cuts and subsidies to chase volume. Volume has returned, but the quality is lower than two years ago.

Incoming CEO Ternus said at the earnings call that he is satisfied with China market performance, with growth mainly from iPhone, followed by Mac. That is true, but the market reaction was a more than 3% after-hours drop in the stock price. Investors care about the $18.75 billion expected shortfall next quarter, not the $17.5 billion year-on-year growth.

Image source: Internet

In China, the difference between 20% and 21% share is not one percentage point, but who sets the rules. Apple still holds high-end pricing power, but it is no longer the Apple that does not need to explain, does not need to cut prices, and does not need to hurry.

Final thoughts

Based on hands-on testing of Apple Intelligence in the already-obtained beta, the China edition of Apple Intelligence looks more like Apple's prudent trial run in China's AI phone race. Its on-device native features focus on practical capabilities such as translation, transcription, and photo processing, while complex creation and deep Q&A are filled in by Qwen. But the AI phone battlefield in China has been fiercely competitive for years, and real-time translation, photo erasure, AI summaries, and smart creation are no longer novel selling points.

Complex creation and deep Q&A are handed to Qwen, and the short-term experience gap can be filled quickly. But there is a long-term hidden concern: once users get used to 'AI capability = Qwen,' the perceived value of Apple's own models will be diluted, and this part of the experience is constrained by cooperation terms and data compliance, not a moat fully controllable by Apple. Domestic manufacturers' agents use their own foundation models plus their own scheduling, while Apple uses 'its own framework + someone else's brain.' The routes differ in substance.

Apple still has a chance to catch up in the time window of domestic manufacturers' agent competition, but the window is narrowing.

References:

First-hand test of Apple China AI Source: Zhidongxi

Everything about Apple is getting more expensive, all because of AI Source: ifanr

- END -