Apple spent two years promising that its AI would be different.
Private, on-device, and woven quietly into the apps you already use.
But as the feature set finally reaches more iPhones, iPads, and Macs, a quieter story is emerging: a large chunk of the installed base is locked out, and the gap is widening with every software update.
Here's the part that rarely makes the keynote.
The full Apple Intelligence suite requires an A17 Pro chip or an M-series processor.
That means anyone still carrying an iPhone 13, 14, or even a base iPhone 15 is watching the marquee feature from the outside.
Apple sold hundreds of millions of those devices.
They are not old by any normal standard, yet they are excluded from the company's biggest pitch of the decade.
The reasons are technical, and Apple has been unusually candid about them.
Running a language model locally demands memory bandwidth and a neural engine that older silicon simply doesn't have.
The company's answer is to offload heavier requests to Private Cloud Compute, a system it says keeps your data encrypted and unreadable even by Apple.
Security researchers have largely praised the design.
But praising an architecture is not the same as shipping features people can actually touch.
Notification summaries, writing tools, and a smarter Siri have arrived in waves, often delayed, sometimes buggy, and occasionally suspended after they summarized a news alert into something misleading.
Users who bought a shiny new iPhone 15 Pro specifically for these tools have spent months waiting for the headline feature to feel finished.
Meanwhile, the competition isn't standing still.
Google has pushed Gemini into nearly every corner of Android, and Samsung has done the same with Galaxy AI, often on hardware that costs less.
Both companies lean on cloud processing, which raises privacy questions Apple loves to highlight.
But there's a tradeoff buried in that contrast: the rival approach works on more phones, today, without a chip lottery.
For American buyers, this creates a genuinely awkward upgrade decision.
If AI features are your reason to buy, the math only works on the newest Pro models or an M-chip Mac.
If you mostly want a fast, reliable phone, the case for spending up gets thinner every month.
Apple is essentially asking customers to pay a premium for a promise that is still being kept in stages.
The European Union's rules have forced Apple to delay some features in that region, and questions about how the company handles AI partnerships keep surfacing.
Every delay feeds the perception that Apple arrived late and is now scrambling to catch up in public.
None of this means Apple Intelligence is a failure.
On-device processing is a real advantage, and the privacy stance could age well as trust in AI erodes.
But the rollout has exposed a hard truth about modern gadgets: the software that sells the phone often can't run on the phone you already own.
Our take: Apple is playing a long game, betting that private AI becomes the standard everyone else has to match.
That bet may pay off, but it's being financed by customers who are paying full price for a feature set that's still under construction.
Final Thoughts
Until the rollout catches up to the marketing, "wait for the next update" is the most honest advice you'll get.