Market Review·September 23, 2026·6 min

Apple’s hardware grew 3.2% in four years. Services grew 59.5%.

We went to check the premium story in the accounts and ran into a problem on the first step: Apple stopped counting the units eight years ago.

Apple at $337.68 on 23 September 2026. Every figure below except the share price comes from Apple's own annual and quarterly filings — the income statement and the product segmentation Apple reports itself; the price is from a market data provider. Where a claim could not be checked against a filing, we say so and leave it out of the argument. Prices move in real time.

There is a story going round about consumer electronics, and it goes like this: people are buying fewer phones and laptops, but manufacturers are doing fine anyway, because the devices cost more. Memory prices are up on AI demand, everything gets passed through, and revenue keeps climbing on falling volumes. Premium, in a word.

It is a tidy story. We went to check it in the accounts, and ran into a problem on the first step.

Apple stopped publishing unit sales in 2018. Proving that fewer units sell at higher prices is hard when the company stopped counting the units eight years ago. Any average selling price for an iPhone you see quoted is an outside firm's estimate, not a filed number. We did not check what the other manufacturers disclose, so we are not going to tell you what they do.

What Apple does publish

Apple breaks revenue into five lines, and the split adds up exactly to the reported total. Four financial years, in billions of dollars:

LineFY2021FY2025Change
iPhone191.97209.59+9.2%
Mac35.1933.71−4.2%
iPad31.8628.02−12.0%
Wearables, home, accessories38.3735.69−7.0%
Services68.43109.16+59.5%
Total365.82416.16+13.8%

Apple's revenue grew by $50.3 billion over those four years. $40.7 billion of it — 80.9% — came from Services. Everything that is a physical object grew by $9.6 billion in total, which on a hardware base of $297 billion is 3.2% over four years.

Not 3.2% a year. 3.2% in total, over four years, without adjusting for inflation. That is not a premium strategy, it is a draw.

And three of the four hardware lines are smaller in dollars than they were in 2021. The Mac has had it worst: it peaked at $40.18bn in FY2022 and now sits at $33.71bn, down 16.1% from its own best year.

The margin, which is the part that looks like the story

Gross margin — what is left of each dollar of sales after the cost of making the thing — has gone one way:

PeriodGross margin
FY202141.8%
FY202344.1%
FY202546.9%
Last four quarters, to 27 June 202648.7%

That is the number the premium story points at, and it is real: nearly seven points of margin since FY2021. But look at what it sits on top of. If rising device prices were doing this, hardware revenue would be rising too. It is flat. What did change, and by a lot, is the mix: Services went from 18.7% of revenue to 26.2%.

We should be careful here, because this is where it would be easy to overclaim. Apple reports products and services gross margin separately in its filings, and we did not have that split in front of us. So: the margin expansion is consistent with the mix shift, and the mix shift is large and documented. Calling it proof would be going further than the numbers we read.

What we could not check, and are therefore not going to assert

The story also covers memory costs rising on AI demand, PC makers pushing upmarket, longer replacement cycles, and a growing second-hand market. We could not pull filings for the other manufacturers, so all of that stays outside this article. It may well be true. It is simply not something we read in a set of accounts, and there is a meaningful difference between the two.

We would rather publish one company verified than five companies assumed. The alternative is how most market commentary gets written, and you can tell.

The binary

Either the premium thesis is right about devices, and hardware revenue starts growing again on price rather than volume — in which case it should show up as hardware dollars, not as a margin percentage, and the next few filings will say so plainly.

Or the growth engine is Services, and the device business is roughly a flat annuity that exists to keep installed base in front of the things that actually grow. In which case the interesting question about Apple is not what an iPhone costs. It is how long Services can compound at this rate, and what happens to it under regulatory pressure — which is a completely different risk from the one the premium story invites you to worry about.

Both at once, no. And four years of filings currently point at the second one.

What this is actually about, if you run a business

You are unlikely to be setting the price of a laptop this quarter. But the mistake in the premium story is one we see in management accounts every single month, and it costs real money.

Total revenue went up, so the strategy is working. That is the sentence. Apple's revenue went up 13.8% and four fifths of it came from one line that had nothing to do with the thing everyone was discussing. Strip the mix out and the headline story evaporates.

The same arithmetic runs in a company doing €4 million: total up 8%, everyone pleased, and underneath it one contract grew 40% while the core business shrank and nobody has looked at margin by line in a year. The number that tells you what to do is never the total. It is the total broken into the pieces that behave differently — which is exactly what Apple's segmentation table does, and exactly what most owner-managed businesses cannot produce on demand.

If you cannot answer "where did the growth come from, by line, and at what margin" in under a minute, you are running the business on the same number the premium story is built on.

The three things we would watch

  1. Hardware dollars, not gross margin. Margin can rise on mix alone. If the premium thesis is right about devices, it has to show up as growing hardware revenue, and so far it has not.
  2. The Mac line specifically. It is 16.1% below its own FY2022 peak. A product that is supposed to be benefiting from a price-led upgrade cycle should not be shrinking in dollars.
  3. Whether Services keeps compounding. It is 26.2% of revenue and provided four fifths of all growth. That concentration is the actual story, and whatever regulatory risk sits on Services is a different question from the one a gadget-pricing debate puts in front of you.

In one sentence

"Over four years Apple's entire hardware business grew 3.2% while Services grew 59.5% — so the premium device story, tested against the only filings we could read, turns out to be mostly a services story in a hardware costume."

Disclaimer. Apple share price $337.68 as of 23 September 2026, from a market data provider. All revenue, gross profit and product segmentation figures are from Apple's own annual (FY2021–FY2025) and quarterly filings, with the trailing-twelve-month figures covering the four quarters to 27 June 2026. Apple has not reported unit sales since fiscal 2018. Claims about other manufacturers, memory pricing, replacement cycles and the second-hand market were not verified against filings and are not asserted here. Prices move in real time. This is general analysis, published identically to every reader — not personalised investment advice and not a recommendation to buy or sell any security. The author is not a licensed investment adviser. Capital at risk. Past performance is not indicative of future results.
We read your company the way we read a listed one. The work behind this analysis — opening a set of accounts and deciding what is possible and what is not — is the same work we do inside private companies. If that is the side that interests you: what a fractional CFO does and what one costs, and the instrument every engagement starts with, the 13-week cash flow forecast.

← All articles

Primus Pilus Capital

Make the money you already have work harder.

Fractional CFO and cash-flow systems: true cost and margin, a rolling 13-week forecast and a management pack your board can use. Fixed fees, never a percentage of your money.

Write to us — no call, no meeting
More articles
Stock Analysis
The $176 Billion Accounting Question Hidden in Big Tech's AI Earnings
Investment Method
The Market Hasn't Changed. What Needs to Change Is How You Read It.
Cash-Flow Systems
The 13-Week Cash Flow Forecast: The Instrument Every SME Should Run On
Stock Analysis
Innodata (INOD): The AI Company That Isn’t a Software Company
Stock Analysis
Dell (DELL): Record AI Earnings — and the Cash Didn’t Move
Stock Analysis
Broadcom (AVGO): Revenue Nearly Doubled, Cash Followed — and the Margin That Rises Is a Mirage
Stock Analysis
Snowflake (SNOW): Six Years Growing 35% a Year — and Not One Cent of Real Profit
Market Review
What 24 Big Managers Are Doing With $7 Trillion
Market Review
We Put 44 Analyst Forecasts Through the Arithmetic. One Cannot Be Met.
Market Review
One Company Crashed 18% Because Nothing Happened — and It Isn't Even Today's Big Story
Stock Analysis
Oracle Has $664 Billion in Signed Orders — and Is Borrowing From Its Own Customers to Deliver Them
Market Review
The Nasdaq Sits 8% Above Its Long-Term Average — and Not ONE of Its 100 Stocks Is Near Its High
Fractional CFO
Fractional CFO: What It Is, When to Hire One, and When You Should Not
Market Review
115 Russell 2000 Stocks Pass the P/E-Below-Growth Screen. Only Twenty Do It Without an Arithmetic Trick.
Market Review
Only 25 S&P 500 Companies Pass the Screen. Nine of Them Are Cases the Method Cannot Judge.
Cash-Flow Systems
The Best 13-Week Cash Flow Forecasting Software — and the Case for Not Buying Any
Market Review
Seven companies passed our screen. In three of them the sum itself was wrong.
Market Review
France added 4 points of debt in a year. Greece took off 9.