The Foldable Is Nice, but Apple's Multiple Hangs on Turning AI Into a Services Line

Apple just reported its strongest June quarter ever — revenue up 16% to $109.4 billion, diluted earnings per share up 29% to $2.02. So when the company takes the stage at its September event and shows off its first foldable iPhone, the natural instinct is to read it as the opening of a new growth cycle. I'd hold that instinct in check. The foldable is a $2,000-plus treat for people who already own an iPhone, and it does almost nothing to answer the question the stock is actually being priced to be paid for.

That question is not about hardware. It is about whether the software layer can keep earning Apple a multiple that says "growth company" while the hardware inside it looks more and more like a mature market. At today's price Apple trades near 35 times trailing earnings and roughly 40 times forward earnings, on a market capitalization around $4.6 trillion. Nobody pays 40 times for a phone. They pay 40 times for the promise that Apple has turned its installed base into a recurring-revenue engine — and, increasingly, that artificial intelligence becomes the next piece of that engine.

The foldable is a niche, not a new era

Settle first what the foldable actually contributes. Supply-chain forecasts peg the device at a price above $2,000, roughly 10 million units in 2026 and 20 to 25 million across 2027. Take even the top of that range at a $2,200 average: about $55 billion a year against a trailing revenue base near $467 billion — roughly a tenth of the top line, and most of it cannibalizing customers who would otherwise have bought a more conventional iPhone in the same upgrade cycle.

That is the twofold problem with the foldable as an investment story. First, it is small enough to be a rounding error in Apple's economics. Second — and this is the part the hardware narrative hides — it shuffles the same users to a higher-priced device within the same installed base. It does not expand the market Apple serves. It is a price-increase dressed as a product category. Against a five-year run of single-digit unit growth in a mature smartphone market, a foldable at 10 million units cannot be the source of Apple's next leg of value creation.

The multiple is a software bet

What actually carries Apple's valuation is Services — subscriptions, app-store fees, search-advertising revenue, cloud storage and the rest. In the June quarter Services brought in $30.7 billion, roughly 28% of revenue, with a paid-subscription base that Apple put at 1.5 billion. That is the asset-light recurring layer that lets a hardware maker command a software multiple. It is also worth noting where the quarter's headline strength came from: iPhone revenue was up 22% to $54.3 billion, but outgoing CEO Tim Cook described recent price increases as driven by a surge in global component costs, made "with considerable reluctance," and Services landed just below the roughly $31.2 billion that analysts had expected. Strip out price, and the volume story in a mature market is closer to the low single digits.

In other words, part of the 16% growth was the company charging more for the same market, not reaching new ones. That matters because the whole bull case is that Apple keeps compounding the software layer without needing to grow the population of devices. The foldable is, at best, a modest upgrade to a single hardware line. It does nothing to the Services trajectory.

AI is still a claim, not a P&L line

Which brings me to what makes Apple genuinely unusual in this AI cycle — and what makes it hard to underwrite at this multiple. Apple is effectively the one large-cap tech company that is not spending hundreds of billions on AI infrastructure. Its capital expenditures over the trailing twelve months came to roughly $10 billion, against a free-cash-flow figure near $137 billion. It runs inference largely on the device, on the largest consumer installed base in tech, and leans on partners like Google for the heavy lifting. By the standards of hyperscalers that are pouring money into data centers, Apple's cost structure for AI is almost comically asset-light.

The flip side is that it has barely converted that advantage into revenue. Apple's AI business — largely Apple Intelligence bundled with iCloud+ subscriptions — is on pace to clear just over $1 billion this year. Against a $467 billion company, that is a rounding error; against a 40-times-earnings price, it is nearly the whole bet. The company introduced a profoundly reworked Siri in June, but the shipping timeline has been vague, and the stock dipped on the day it was announced — the market read it correctly as promise, not delivery.

This is the divide that separates Apple's position from its price. It has the moat — the installed base, the on-device model, the low cost of serving AI to a billion-plus users. What it does not yet have is the operating result: a Services-style line that shows AI turning from a free feature that defends hardware upgrades into recurring revenue that justifies 40 times earnings. That is a claim in search of proof, and the foldable proves nothing about it.

So do not mistake the foldable for the thesis. It is a hardware ceiling, not the engine of the multiple. The stock is a wager that Apple does for AI what it did for subscriptions a decade ago — turn a feature into a margin-rich, recurring business on an installed base no one else can match. The chips that decide that bet are already on the table: can Apple show, quarter after quarter, in the Services line, that AI earns its keep the way App Store fees once did. Until that appears in the numbers, Apple is an expensive software promise wearing a very nice new phone on its face.