Three new iPhones arrive in September, and not one of them sits below the Pro tier [15], so the plain handsets that get bought by the pallet are not in the room.
Teams tell themselves the autumn keynote is when the range resets and the price list gets a new bottom rung. What a buyer with 400 seats does with the keynote is narrower: look up the two cheapest part numbers, confirm what last year's model still costs as a carryover, then place the order. This year the first check comes back empty. The second is the fact that decides how much this actually costs, and Apple has confirmed only that the event is happening [4], which leaves the carryover list unknown and the foldable's price with it [10].
The reason a hardware event is carrying this much weight is that the software question got answered by contract. Siri was rebuilt on a custom Google Gemini model and shipped as a standalone app, routing queries between on-device models, Private Cloud Compute and a 1.2-trillion-parameter system Apple did not build [7]. Apple settled a $250m class action over AI features marketed before they were ready [8], and according to TNW the Pixel 11 had Gemini in the operating system weeks before Siri got comparable models [9]. Ternus, who joined Apple in 2001 on the product design team and ran hardware engineering from January 2021 [6], is answering with shapes instead: Mark Gurman reported for Bloomberg that this is the start of the biggest run of device releases in the company's history [3], and the foldable is described as a book-fold with a roughly 7.6-inch inner display and a 5.5-inch cover screen on 2nm A20 Pro silicon, none of which Apple has confirmed [10].
TNW sets out the fair rebuttal, and it holds up: hardware including silicon is where Apple's advantage is most defensible, and licensing a frontier model while building differentiated devices is a rational division of labour rather than a retreat [13]. Apple's third-generation foundation models are a real programme, and on-device inference is a genuine strategic position [14]. Both of those can be true while the buyer still holds a mix problem. Crease visibility isn't a metric that decides fleet standardisation, yet it's the axis Samsung has been working for the seven years since the first Galaxy Fold, most recently with the Z Fold8 Ultra in July at just over four millimetres unfolded [11]. Foldables have stayed a premium niche for that whole run [12].
The arithmetic that settles it is small. Take the units due for replacement before the spring window opens [16] and multiply by the gap between what a plain handset cost last cycle and what a Pro costs now. Set that against the cost of running the current fleet two extra quarters: battery replacements, warranty extensions, the handsets that were already going to fail. If the premium is larger, September becomes a line in the file for spring rather than a purchase. If it is smaller, buy Pro and stop spending maintenance money on hardware that has already left the plan.
The open question is whether the previous generation stays on sale at a lower price, and that is a question for a reseller before the keynote rather than after. For anyone buying in volume, the decision has moved to spring 2027 [5], and the only thing 9 September settles is how long the fleet already in people's pockets has to last.
