iOS 27: What Actually Changes on Your iPhone - And What the Marketing Is Overselling

Every September, Apple releases a new iOS version. Every June, the coverage that precedes it reads roughly the same: a long list of every announced feature, organised by app, with equal enthusiasm applied to transformative changes and to minor refinements that most users will never consciously notice.

iOS 27 deserves better than that treatment, because it is genuinely a more consequential update than most - and at the same time, it has one significant caveat baked into it that the marketing carefully avoids dwelling on, and that anyone reading a feature list without context will almost certainly miss.

The update ships to the public in September 2026, arriving pre-installed on the iPhone 18 Pro lineup. Developer beta is already running. Public beta arrives in July. Here is what is actually worth understanding, what the performance improvements mean in real numbers, and why the Siri AI upgrade - the centrepiece of the entire release - is a more complicated story than "Siri is finally good."

What Actually Matters: The Siri Overhaul

Start here because everything else in the release sits downstream of this decision.

Apple has rebuilt Siri from the ground up. The new version, called Siri AI, is built on the next generation of Apple Foundation Models developed in collaboration with Google, running on Gemini infrastructure at the highest capability tier. This is not an incremental improvement to the Siri that confused people by misunderstanding "call John" for a decade. It is a genuinely new assistant that happens to share a name with the old one.

The functional differences are substantial. Siri AI holds full back-and-forth conversations with persistent context - you can ask a follow-up question about something you said three exchanges ago and it will understand what you are referring to without re-stating the whole thing. It searches across Mail, Messages, Notes, Calendar, Reminders, and third-party apps that have adopted Apple Intelligence APIs to answer questions about your own life - "when did I last hear from my accountant" or "where did I save that restaurant recommendation" are now genuinely answerable. It generates written content, analyses files, searches the web, and handles complex multi-step tasks through a new dedicated Siri app that syncs conversation history across all your Apple devices through Private Cloud Compute.

The camera integration is the feature that will get the most real-world daily use despite being one of the less-headlined additions. Siri Mode in Camera - previously called Visual Intelligence, now renamed to make it more discoverable - lets you point your camera at anything and ask questions about it in a natural conversation that gets saved in the Siri app. Point it at a restaurant and ask about the menu. Point it at a plant and ask how to care for it. Point it at nutritional information and ask whether it fits your diet. The renaming from Visual Intelligence to Siri Mode is itself a signal: Apple is explicitly positioning this as Google Lens territory, which is an acknowledgment that Google's implementation of the same concept has been genuinely useful while Apple's equivalent was underused partly because people did not know it existed.

The dedicated Siri app is a structural change worth noting separately from the capability improvements. Having a conversation history that persists, that you can return to, that syncs across your iPhone, iPad, Mac, and Apple Watch, moves Siri from something you invoke and forget toward something closer to an ongoing working relationship with an assistant that knows what you have asked before. Whether that framing makes you enthusiastic or mildly unsettled probably depends on your existing relationship with the concept of AI assistants.

The Performance Numbers Are Real, and They Are Larger Than You Would Expect

Buried behind the Siri coverage are performance improvements that the official iOS 27 page quantifies with unusual precision for Apple: app launches up to 30% faster, new photos loading in the Photos library up to 70% faster, AirDrop transfers up to 80% faster.

These are not marketing-speak percentage improvements that in practice manifest as "slightly less slow." For the app launch and photo loading numbers specifically, the improvements are measurable in the use cases where they matter most - opening a camera quickly enough to catch a moment, or scanning your library after a day of shooting without watching a spinner.

The 80% faster AirDrop figure is the one that will get the most appreciative response from people who regularly transfer files between Apple devices. AirDrop's inconsistency has been a persistent frustration - the transfer that completes instantly, and the identical transfer that inexplicably takes 45 seconds and then fails - and an 80% improvement in transfer speed, if it holds across real-world conditions rather than just the controlled conditions a spec claim implies, would materially change how useful the feature is for anything involving large files.

App launch performance at 30% faster is measured from the iPhone 18 Pro hardware - on older devices in the iOS 27 compatibility range, the improvement will be smaller. iOS 27 is compatible with the iPhone 11 and later, which means Apple is shipping this software to devices up to six years old. The performance improvements scale with the underlying hardware; an iPhone 11 will not hit the same numbers as an iPhone 17 Pro.

The Wi-Fi and cellular network switching improvement is the one that sounds boring until you have been on the receiving end of the failure mode it fixes. The new system more seamlessly chooses between available connections - switching from home Wi-Fi to cellular as you walk out the door, or from cellular to airport Wi-Fi as you land, without the 30-second limbo where your phone is connected to one network but routing through another. For FaceTime calls and navigation specifically, this matters in exactly the moments where you most need the connection to be reliable.

The New Features Worth Actually Using

Moving through the individual app updates with an honest filter applied - what will the median iPhone user notice and use, versus what will sit untouched in a settings menu somewhere:

Calendar's natural language event creation is the addition that will save the most cumulative time for the most people. The ability to type "coffee with Sarah at noon next Thursday" and have the event created correctly without navigating date pickers and time selectors is the kind of small improvement that compounds across hundreds of interactions over a year. The implementation requires a specific syntax for dates that are not the current one - you need to type the date explicitly and tap the suggestion, rather than expecting it to parse "Thursday" as meaning next Thursday if today is Monday - which is worth knowing before assuming it is as seamless as the marketing implies.

Safari's tab organisation is more useful than it initially sounds. The AI grouping of related tabs - and specifically the Notify Me feature that lets you flag a tab and have Apple Intelligence alert you when the page updates - addresses the actual workflow problem of keeping a research session alive without leaving a browser window open indefinitely. Tab groups have existed in Safari for years; the AI-driven organisation that adds related tabs automatically and lets you save a research thread as a named group is the layer that finally makes the feature feel like it was designed for how people actually use their browser rather than for a theoretical organised person who proactively manages their tab situation.

Photos loading speed deserves its own mention separate from the performance section because it affects a specific daily interaction. Opening the Photos app, having thumbnails appear instantly rather than loading progressively, makes browsing and sharing feel meaningfully different. For iPhone users with large libraries - which after several years of high-resolution shooting means most people - this is one of the improvements that will be noticed in the first week and appreciated ongoing.

Shared Albums with Android and Windows support is the quiet practical win for anyone who shares photos with people outside the Apple ecosystem. The ability for non-iPhone users to contribute their own photos through iCloud.com - rather than requiring them to use a third-party app or go through a separate sharing workflow - removes the most common friction point in shared family or event albums. Full-resolution sharing is specifically listed as a new capability, which closes the long-standing limitation where shared albums compressed photos in ways that mattered for anyone who wanted to print or archive them.

RCS improvements for Android conversations - specifically the ability to reply to a specific message in an RCS conversation, and proper display of reactions on images and videos - continue iOS 27's incremental but consistent narrowing of the gap between iMessage functionality and what iPhone users experience when texting Android contacts. These are not headline features. They are quality-of-life improvements that matter considerably to the people who have spent years watching their reaction emoji appear as "[loved a photo]" on the other end.

The Features That Sound More Exciting Than They Are

Honest assessment requires calling out the additions that the feature list presents as meaningful and that are, in practice, more limited than the framing implies.

Write with Siri system-wide - the ability to have Siri generate or proofread text in Notes, Mail, Messages, and other apps - is presented as a significant writing tool and will get significant use from some people. The limitation worth naming: matching your writing style, punctuation, and tone, as Apple describes it, requires enough of your prior writing in the relevant app to have something to learn from. For a Notes app used primarily for short reminders or a Messages conversation that is mostly brief replies, the style-matching feature has very little to work with and will produce generic output. The capability is real; the "sounds like you" promise requires conditions that many users will not have.

Wallet Insights - the new feature visible in iOS 27 Beta 2 that connects accounts to show spending patterns, recurring transactions, and balances - is genuinely not finished. Beta 2 shows a splash screen for the feature with the functionality labelled as coming, not currently operational. This is normal for early betas; noting it here because any coverage written from the beta that presents Wallet Insights as a ready feature is getting ahead of where the software actually is.

Liquid Glass slider - the system-wide customisation control for the Liquid Glass design introduced in iOS 26, letting you adjust the effect from ultraclear to fully tinted - is a welcome addition for people who found the default level of the effect either too subtle or too overwhelming. It is also a very minor thing, and its presence in marketing materials as a notable feature reflects more about how the Liquid Glass design has divided opinion than about the intrinsic significance of a transparency slider.

The Caveat That Changes Everything for Some Readers

This is the part of the iOS 27 story that belongs in every review and is present in almost none of the breathless WWDC coverage: Siri AI, the feature that defines the entire release and is the reason Apple described iOS 27 as its most ambitious update in years, will not be available in the European Union at launch.

The reason is the Digital Markets Act. Apple and the EU have not reached agreement on the specific terms under which the deeply personal, cross-app data access that Siri AI requires can be implemented in compliance with European interoperability and gatekeeping requirements. Apple has indicated it expects to bring Siri AI to EU users eventually, but has given no timeline, and the history of Apple's DMA negotiations suggests that "eventually" has consistently meant longer than Apple's initial communications implied.

Siri AI is also not available in China at launch, due to separate regulatory requirements around data processing and AI features.

Morgan Stanley's estimate is that the EU and China together account for roughly 35% of iPhone shipments. An iOS update whose defining feature is unavailable to over a third of the people who will install it is a meaningful qualification that the feature list format of most iOS 27 coverage does not capture.

For readers in the EU specifically: iOS 27 is still worth installing for the performance improvements, the RCS updates, the Camera and Safari improvements, and the full set of Apple Intelligence features that do not involve Siri AI's deep personal context access. It is not, however, the release that makes your iPhone feel as different as Apple's marketing implies - that version of the update is coming later, on a timeline that is not publicly committed.

What iOS 27 Actually Is, in One Honest Summary

iOS 27 is the update where Apple finally made Siri into something it should have been years ago, shipped it at a time when the hardware can actually handle running a capable AI model, and had to simultaneously acknowledge that two of its most important markets cannot have the centrepiece feature at launch.

The performance improvements are real and will be felt. The camera integration is genuinely useful. The underlying engineering work - the rebuilt Siri architecture, the Private Cloud Compute infrastructure, the Apple-Google Foundation Model collaboration - is significant regardless of whether the output lands perfectly in beta 2 or takes a few more months to mature.

The public beta in July is when most people will actually form their first impression of what iOS 27 feels like to use daily - not from keynote demos, but from putting Siri AI through the specific requests and frustrations of their own life and seeing where it helps and where it still falls short.

That is the more honest version of what iOS 27 is: a platform shift in how the iPhone's core assistant works, built on genuinely different underlying infrastructure than its predecessor, with the realistic expectation that a complete, polished version of what Apple announced in June will take until late 2026 or into 2027 to fully materialise for all the people who own an iPhone.

That is still worth being interested in. It is just not the "Siri is fixed and ready right now" story some of the coverage implied.