Everything About AI
NextFin News - Apple delivered a record September quarter but told investors to expect thinner profits through the holidays, and the market punished the disconnect. The iPhone maker posted fiscal fourth-quarter revenue of $114.48 billion, up 12% from a year earlier and above the $113.47 billion consensus, with iPhone sales jumping 17% to $57.58 billion and diluted earnings per share rising 14% to $2.10. Yet shares fell 2.05% to $331.37 on November 3 after the company guided December-quarter gross margin to 45%–46%, below the 46.7%–47.5% range analysts had expected, and flagged what CEO John Ternus called “unprecedented” memory costs. All market figures are as of the November 3 close.
The tension is stark: Apple’s top line is accelerating while the profit that falls through to the bottom line is being squeezed. In the first quarterly report under Ternus, who took over from Tim Cook effective September 1, 2026, the company is asking investors to look past a near-term margin trough that management says is driven by a once-in-a-cycle component shortage rather than any loss of pricing power. The question investors are left with is whether this is a cyclical dip to buy through or the first crack in a margin model that has powered Apple’s valuation for a decade.
The Numbers: A Beat on Demand, a Miss on Profit Quality
The headline figures were strong enough to satisfy a growth-focused investor. Total revenue set a September-quarter record, products revenue rose 12% to $82.9 billion, and services revenue climbed 10% to an all-time high of $31.6 billion. Net income reached $30.7 billion, also a September-quarter record, and gross margin of 47.9% looked healthy on its face. For the full fiscal year, revenue reached $478.8 billion, up 15%, with diluted EPS of $8.98, up 20%.
But that 47.9% carries an asterisk the size of the holiday quarter itself. Roughly one percentage point of the margin came from tariff refunds — a one-time benefit that will not repeat. Strip it out and the underlying margin sits closer to 46.9%, which is essentially flat against the 46.9% gross margin Apple reported for fiscal 2025 and well below the 48.2% the company posted in the December 2025 quarter. That makes the 45%–46% guide for the December quarter look less like a temporary dip and more like a continuation of a pressure trend that began before Ternus ever stepped to the podium.
The December outlook was otherwise encouraging. Apple expects revenue growth of 7%–10% year over year, which could set another quarterly sales record, and services growth of 8%–9%. Operating expenses are guided to $20.4 billion–$20.8 billion, up from $19.2 billion in the September quarter, reflecting continued investment in artificial intelligence and the Siri cloud buildout. Other income and expense is expected to be roughly $300 million, with a tax rate near 17%.
So the market was not reacting to weak demand. It was reacting to the composition of the profit. When a company beats on revenue and earnings per share but warns that the engine of its valuation — the high-margin services layer and the pricing discipline of its hardware — is about to compress, investors recalculate the multiple they are willing to pay. The 2.05% decline on November 3, which followed a 1.27% drop on November 2, leaves the stock roughly 5% below its October 29 close of $348.91 and about 5.8% under its 52-week high of $351.73.
The Mechanism: Why Memory Costs Are Different This Time
The official explanation is a supply shock, and management laid it out with unusual precision. CFO Kevan Parekh told analysts that the midpoint of the December-quarter margin guide implies a sequential decline of roughly 140 basis points once the September quarter’s tariff-refund benefit is excluded, and that most of that decline is driven by higher memory costs.
“Memory pricing is unprecedented,” Ternus said on the earnings call, adding that Apple raised prices on Mac and iPad in June and on iPhone in September to partially offset the cost pressure.
This is not a normal inventory cycle. In a typical downcycle, component prices fall and assemblers benefit from cheaper inputs. What Apple is describing is the opposite: a memory shortage severe enough that the world’s largest buyer is stockpiling parts and still cannot get enough. Chief Operating Officer Sabih Khan said manufacturing purchase obligations rose $33 billion from the June quarter to the September quarter — a defensive move to secure scarce components. That Apple is willing to raise prices, a rare step for a company that normally absorbs cost pressure rather than passing it through, is itself evidence of how tight supply has become.
The transmission channel from memory costs to Apple’s margin is mechanical and hard to dispute. Memory is a direct input into every iPhone, Mac, and iPad. When the cost of that input rises faster than the company can raise selling prices, gross margin compresses dollar-for-dollar. The 140-basis-point sequential decline Parekh outlined is not a mix-shift story or a one-time charge; it is a cost-of-goods-sold story, and those are the hardest margins to defend because they require either supplier relief or price increases that risk unit volume.
There is a second cost layer buried in the guidance that investors should not miss: the Siri artificial-intelligence buildout. Parekh noted that Siri AI cloud costs will flow through services cost of sales, and that a portion of device revenue tied to AI features is deferred over the life of the device. In plain terms, Apple is spending on AI infrastructure today while recognizing some of the revenue later. That timing mismatch is a deliberate investment, but it is also a margin headwind that will persist well beyond the memory cycle. The services segment, which carried a 76% gross margin in the September quarter — up 40 basis points sequentially — is the profit engine that funds everything else; any structural erosion there matters more than a hardware blip.
The Second-Order Effect: From Margin Compression to Multiple Compression
The first-order effect of rising memory costs is obvious: lower gross margin. The second-order effect is what actually moved the stock. Apple’s valuation premium has rested on the assumption that its margin structure is durable — that the shift toward services creates a floor under profitability even when hardware cycles turn. A 140-basis-point sequential margin decline, guided with confidence by management itself, forces analysts to rebuild their models on a lower terminal margin rate. That is multiple compression, and it hits harder than a single-quarter earnings miss.
The propagation chain runs further than Apple’s own share price. Apple is the marginal buyer in the memory market, and its $33 billion increase in purchase obligations signals that the shortage is industry-wide. That means the same cost pressure is landing on Dell, HP, Lenovo, and every other device maker that cannot command Apple’s pricing power. The difference is that those companies lack a 76%-margin services business to absorb the shock. If Apple — the best-positioned hardware vendor in the world — is being forced to raise prices and still sees margins compress, the smaller players face a steeper climb. The risk is not confined to one stock; it is a sector-wide margin story wearing an Apple ticker.
There is also a cross-asset read. A durable step-down in Apple’s margin rate would reduce the earnings growth that has justified its premium valuation multiple. With the stock trading near $331 and a market capitalization of $4.84 trillion, even a small change in the assumed terminal margin translates into hundreds of billions of dollars in equity value. The bond market is watching too: Apple’s $80 billion debt load is serviced comfortably by its $28.1 billion in quarterly operating cash flow, but a lower margin trajectory would narrow the cushion that supports its investment-grade credit spreads.
Cyclical or Structural: The Call That Determines the Stock
Here is the judgment that matters, and it requires separating two forces that management presented as one. The memory shortage is cyclical — it will revert. Component cycles are mean-reverting by nature: high prices trigger capacity expansion, and expanded capacity eventually breaks prices. Apple’s $33 billion increase in purchase obligations is itself evidence that supply is being pulled forward, and the company’s willingness to lock in supply suggests it expects the shortage to ease as new fabrication capacity comes online.
But the AI cost structure is structural — it will not revert on its own. Every query processed through Apple’s cloud, every Siri feature running with server assist, adds a permanent line item to the cost base. Unlike memory, which is a commodity with a well-understood cycle, AI inference costs are tied to usage growth: the more successful Apple Intelligence becomes, the more it costs to run. This is a regime change in Apple’s economics, from a business that scaled services with near-zero marginal cost to one where usage growth carries a real marginal cost.
The market is currently pricing the cyclical leg and underpricing the structural leg. A 140-basis-point margin hit from memory is recoverable; a permanent 100–200 basis-point step-down in the services margin rate as AI scales is not. That is why the stock’s reaction is better understood as a repricing of the terminal margin assumption than as a vote against holiday demand. The contrast between the 76% services margin and the 37.2% products margin — and the fact that AI costs land on the services side of the ledger — is the crux of the matter.
The Counter-Thesis: This Is Just a Timing Mismatch
The strongest case against this read is straightforward: Apple has been here before, and it has always recovered. The company raised prices in June and September specifically to offset these costs, which means the margin compression is a lagging indicator — the price increases have not yet flowed through the revenue line. Once they do, margins should rebound even if memory costs stay elevated. Moreover, the installed base just crossed 2.5 billion active devices and paid subscriptions topped 1.5 billion, giving Apple a services foundation that no competitor can match. If the holiday quarter delivers the 7%–10% revenue growth management expects, the narrative shifts back to growth, and the margin miss becomes a footnote.
This argument is credible, and it is the base case embedded in the current price. But it rests on two assumptions that the earnings call did not fully support. First, it assumes price increases will not damage unit demand — and Apple has not yet disclosed whether the September iPhone price hike affected upgrade rates. Second, it assumes the AI cost ramp is a temporary investment phase rather than a new steady state. If either assumption breaks, the recovery takes longer and runs shallower than the consensus expects.
The falsifying signal is specific: if Apple’s March-quarter 2027 gross margin does not recover to at least 47% — the level of the September quarter excluding the tariff benefit — the cyclical-recovery thesis is wrong. That would indicate that AI infrastructure costs, not memory, are the dominant margin driver, and the market would be right to assign a lower multiple.
What Comes Next: Three Horizons
Short term (the holiday quarter): Expect volatility. The 45%–46% margin guide is a known headwind, and any sign that iPhone 18 or iPhone Duo supply constraints are easing would be read as a positive. The revenue guide of 7%–10% growth gives management room to beat, which could stabilize the shares even if margins lag. The dividend of $0.27 per share, payable November 16 to holders of record November 12, provides a small floor for income-focused holders.
Medium term (through fiscal 2027): The key variable is memory pricing. If component costs normalize as new capacity comes online, margins should recover toward 48% and the stock re-rates higher. The watch item is the price-increase pass-through: Apple needs to show that higher selling prices are not eroding the installed-base upgrade cadence. The company’s $137 billion in cash and marketable securities, against $80 billion of total debt, gives it ample room to absorb cost pressure or accelerate buybacks if the shares stay weak — it returned $28 billion to shareholders in the September quarter alone, including $24 billion in repurchases.
Long term (structural): This is where the real risk and opportunity sit. If Apple can scale Apple Intelligence while keeping services margins above 70%, the company proves that AI can be monetized without sacrificing profitability — a template that would support a premium valuation. If AI costs consume a growing share of services revenue, the multiple compresses regardless of how well the hardware cycle performs.
The base case is a cyclical margin trough followed by a recovery, with the stock range-bound until the March quarter confirms the inflection. The upside case is memory relief arriving faster than expected, lifting margins back above 48% by mid-2027. The downside case is AI infrastructure costs proving stickier than management anticipates, turning a 140-basis-point dip into a permanent step-down.
Apple’s holiday forecast is upbeat, but the market is no longer paying for holiday revenue — it is paying for the margin that revenue produces. Ternus’s first act as CEO has been to trade near-term profit quality for long-term positioning, and the next two quarters will tell investors whether that trade was priced correctly or whether the era of Apple’s structural margin expansion has quietly ended.
Explore more exclusive insights at nextfin.ai.
