Apple’s Q3 FY2026: $109.4 Billion in Revenue, a 50% Gross Margin, and $6.8 Billion in Nine-Month Capex — The Anti-Hyperscaler Thesis, Quantified

In the same earnings week that Amazon, Alphabet, and Meta each spent more on capital expenditure in a single quarter than Apple spent in nine months, Apple posted its best June quarter ever — and returned more cash to shareholders than it invested in infrastructure.

What Happened

Apple reported fiscal Q3 FY2026 results on July 30, 2026, covering the quarter ended June 27. Every headline metric set a June-quarter record. Total revenue reached $109.4 billion, up 16% year over year. Net income was $29.8 billion. Diluted EPS came in at $2.02, up 29% — though Apple’s own press release flags that tariff refunds contributed approximately $0.11 to EPS and roughly two percentage points to the 50.1% gross margin. Strip those out and the underlying margin sits closer to 48%, still exceptional, but the distinction matters for modeling the run rate.

By segment: iPhone revenue rose 22% to $54.3 billion; Mac jumped 29% to $10.4 billion; Services grew 12% to a record $30.7 billion. Greater China, a geography that has absorbed significant geopolitical and competitive pressure, grew 22%. One outlier: iPad revenue fell 6%, the single segment that missed the record sweep. R&D spending surged 32% to $11.7 billion, a figure Tim Cook contextualized with a reference to the “all-new Siri AI” previewed at WWDC26 and an installed base described as being at an all-time high. Over the nine months ended June 2026, Apple generated $117 billion in operating cash flow, returned approximately $62 billion via buybacks, and paid $11.8 billion in dividends.

The number Apple barely mentions — and the one that defines the structural story of this earnings week — is capital expenditure. Over those same nine months, Apple spent approximately $6.8 billion in capex.

The Structural Read

The honest framing first: the capex comparison is partly structural, not merely a product of capital discipline. Amazon, Alphabet, and Meta operate hyperscale cloud businesses — AWS, Google Cloud, Azure adjacents — that sell compute to third parties at industrial scale. Apple does not. It does not need fifty-billion-dollar-a-quarter data-center budgets because it is not in the business of renting inference capacity to the world. To compare their capex lines as if they reflect equivalent strategic choices would be a category error.

That caveat noted, the contrast is still structurally significant. Apple is investing in AI — R&D up 32% to $11.7 billion is not a rounding error — but through a fundamentally different mechanism: R&D and rented compute rather than owned infrastructure. Its model is to monetize the AI wave at the distribution endpoint: 2+ billion active devices, a Services business compounding at $30.7 billion per quarter, and a hardware refresh cycle that Apple Intelligence will eventually accelerate. The installed base is the moat. The capex line reflects a bet that you can win the AI era as a Distributor rather than as a Builder of the underlying infrastructure.

The mirror image of this thesis is playing out simultaneously across the earnings week. Meta’s free cash flow fell to $784 million as capex consumed nearly all operating cash generation. Alphabet’s free cash flow turned negative. These are deliberate bets — pre-buying capacity for a demand curve they believe is certain — but they are bets with real short-term cost. Apple, by contrast, generated $117 billion in operating cash flow over nine months, returned $73.8 billion to shareholders, and still posted a 50% gross margin. The $5 trillion anti-capex distribution-endpoint thesis is not a contrarian take anymore — it is what the income statement says.

One genuine risk to the thesis: Apple Intelligence and the “all-new Siri” shown at WWDC26 are still ramping. The “Apple is behind on AI” narrative has not been retired by this quarter’s results — it has only been set aside by strong hardware and Services numbers. If on-device AI fails to drive upgrade cycles, and if third-party LLM integrations become the primary AI experience on Apple devices, the distribution-endpoint moat narrows. The 6% decline in iPad revenue is a reminder that not every category bends to the installed-base flywheel on command.

Three Implications