Apple may not be bringing Apple Intelligence to iPhone 15, which has 6GB of RAM vs. iPhone 15 Pro's 8GB, due to RAM requirements for on-device AI processing
Older iPhones might well be able to cope with Apple's AI, so why won't your $900 year-old model run it? Here's why.
Context & Ripple Effects
The reported divide follows a hardware split already visible in the iPhone 15 generation: iPhone 15 and 15 Plus were identified with 6GB of RAM, versus 8GB in the Pro models. The story matters because it turns a largely invisible specification into a potential boundary for a flagship software capability.
Later coverage raises the apparent threshold further, reporting that Apple's most powerful on-device model requires devices with 12GB or more of RAM. That makes the iPhone 15 question an early example of AI features being gated by local compute capacity rather than device age alone.
First-order effects
- If the reported requirement holds, iPhone 15 and 15 Plus owners would be excluded from Apple Intelligence capabilities available on iPhone 15 Pro models, despite being in the same product generation.
- Apple's Pro hardware tier gains a concrete functional distinction: its additional RAM would support on-device AI workloads that the 6GB models may not accommodate.
Second-order effects
- RAM capacity becomes more salient in iPhone purchase and upgrade decisions, giving Apple an incentive to communicate AI compatibility by device tier rather than treating software updates as broadly uniform.
- The split pressures handset makers to balance memory costs against AI feature eligibility, as local AI performance and availability increasingly depend on hardware configuration.
Third-order effects
- If on-device AI keeps raising memory requirements, software support windows may increasingly be defined by compute headroom, producing more fragmented feature access within nominally current device lineups.
- The pattern points to heterogeneous AI compute: vendors may reserve the most capable private, local models for hardware with sufficient memory while offering different feature sets elsewhere.
The trend: Consumer AI is shifting device differentiation toward the memory and compute capacity required to run models locally.