Apple Cuts iPhone 18 Pro Orders as AI Memory Costs Reach the Handset
Apple has told suppliers to cut component orders for the iPhone 18 Pro and Pro Max by 15 to 20 percent for October, according to a Nikkei exclusive, citing soft demand after AI-driven memory costs pushed prices up. The same memory squeeze cut global smartphone shipments to their lowest third quarter since 2012, with over 40 percent of models seeing price hikes this year.
On this page
The order cut, in the supplier's words
Apple has told suppliers to reduce component orders for the iPhone 18 Pro and iPhone 18 Pro Max by 15 to 20 percent for October, according to a Nikkei exclusive. The supplier quote in the report is unusually direct: "In October alone, we are seeing orders from Apple reducing by 15% to 20% for both the premium models; we don't know how things would develop." The stated cause is soft demand, driven in meaningful part by higher prices, which were pushed up by surging memory chip costs across the industry.
The memory cost surge is not an Apple-specific problem; it is the demand curve of AI data centers arriving in the consumer supply chain. Data center buyers are outbidding phone makers for DRAM and NAND, and the phone industry is absorbing the repricing.
The wider memory crunch, by the numbers
The Q3 data puts the squeeze in context: global smartphone shipments fell roughly 20 percent year over year to their lowest third-quarter level since 2012, and full-year projections for 2026 sit around a 13 to 17 percent decline to just over a billion units. Rest of World's reporting found more than 40 percent of smartphone models saw price hikes in 2026, with emerging markets hit hardest, worsening the digital divide that cheap Android hardware had been closing.
The beneficiaries are the memory makers themselves: Samsung's profit hit a record on the AI memory boom, and the analysts' expectation is that chipmakers keep prioritizing data center demand through the shortage, with mid-range phone prices forecast to rise further into 2027. For anyone watching on-device AI specifically, the irony is structural: the same AI boom funding bigger cloud models is making the hardware that could run local models scarcer and dearer.
What it means for on-device AI's roadmap
Two consequences follow for the local-AI roadmap. First, the flagship tier where big on-device models debut gets more expensive and more supply-constrained in the near term, which slows the demo-to-default pipeline for running large models on phones. Second, efficiency stops being a nice-to-have: quantization, smaller embedders, expert-caching, and models like the 740M-class multimodal embedders shipping this month are how local AI fits into devices whose memory allocations are about to get tighter, not looser.
The suppliers, notably, do not know how October develops, and neither does anyone else; the order cut is a signal, not a trend line. But the direction of the constraint is clear, and it runs from the AI data center straight into the price tag of the next phone.