Apple’s iOS 27 Siri Gets Liquid Glass and a Google Brain

Apple’s iOS 27 Siri will spring out of the Dynamic Island wearing a translucent “Liquid Glass” coat when the company opens its Worldwide Developers Conference (WWDC, Apple’s annual software event) keynote on June 8, according to renderings published by Bloomberg reporter Mark Gurman. The redesign hands the assistant a standalone app, a system-wide search bar, and rich result cards that fan out of the notch.

Underneath that glass runs a custom Google model carrying 1.2 trillion parameters, roughly eight times the size of the cloud model Apple built in-house, and Apple is paying about $1 billion a year for it. For a company that spent a decade insisting it owned its own silicon and software top to bottom, that is the part worth reading twice.

What the Leaked Renderings Show

Gurman’s images landed 11 days before the keynote, and they confirm the deepest cosmetic change to the assistant since it launched in 2011. Gone is the glowing rainbow ring that wraps the bottom edge of the screen. In its place, a pill-shaped animation pulses inside the Dynamic Island, with a soft “searching” label while the assistant works through a request.

The muscle memory shifts too. Here is what the reporting describes:

• Swiping down from the top center of the screen now opens a “Search or Ask” panel for typing or speaking a question.
• Notification Center moves to a swipe from the top-left corner, because the assistant has claimed the central real estate.
• Pulling down further on the search panel launches a full chatbot-style conversation window.
• A dedicated Siri app stores past conversations and accepts file, photo, and document uploads for analysis.
• Answers surface as translucent cards that pop out of the Dynamic Island rather than filling a separate screen.

If that layout sounds familiar, it should. The standalone app borrows its shape directly from OpenAI’s ChatGPT and Anthropic’s Claude, down to the scrolling history and the message-thread aesthetic. Apple’s own coverage of the new standalone Siri chatbot and business app strategy points to the same dual-app split.

The Google Brain Behind the Glass

The visual makeover is the easy headline. The engineering decision underneath is the one that breaks with two decades of Apple habit. After the company’s own foundation models came up short, Apple licensed a fully trained version of Google’s Gemini model family to do the heavy reasoning, in a deal the two firms confirmed in January.

Eight Times the Brain

The custom model Apple is renting is not the consumer Gemini you can open in a browser. It is a bespoke build tuned for Siri and Apple Intelligence, and the size gap with Apple’s in-house cloud system is stark.

Renting, Not Sending

The crucial detail for privacy-minded buyers: the bigger model runs inside Apple’s Private Cloud Compute security architecture, not on Google Cloud. Apple licenses the trained weights and serves inference on its own hardware-isolated servers, so queries are processed and then discarded rather than handed to Google. The arrangement is closer to leasing an engine and dropping it into your own chassis than outsourcing the whole drive.

That nuance matters because the deal still reads, to the wider market, as Apple conceding the most important contest in consumer technology. Our earlier look at how Gemini quietly took over core iPhone features traced the same shift from a product Apple wanted to own to one it now buys.

Why Apple Blinked

This is the bill arriving for a promise made in 2024. That year Apple stood on stage and pitched a context-aware Siri that could act across your apps, then spent the next year and a half quietly pushing the date as its own models failed to deliver the goods. The personalized assistant slipped, then slipped again.

By the start of 2026, the company chose pragmatism over pride. Rather than keep waiting on internal silicon brains, it went shopping. On the company’s first-quarter earnings call, chief executive Tim Cook framed the Google tie-up not as a retreat but as a foundation Apple was comfortable building on.

We are very happy with the collaboration with Google.

That line, from Cook, lands differently when you remember that Apple positioned Google’s search payments and Android as existential rivals for years. The assistant that was supposed to prove Apple could win AI on its own terms now leans on the company it competes with hardest.

Siri Takes Over the Camera

The redesign reaches well past the home screen. The Camera and Photos apps inherit a new layer of AI tools, with a fresh mode set to replace the current Visual Intelligence feature that reads what the lens sees.

Point the camera at an object and you will be able to run it through a Google reverse image search or bounce the file to third-party AI agents for instant analysis. The editing suite picks up two tools aimed squarely at the generative crowd:

• Extend generates visual detail beyond a photo’s original borders, filling in scenery the frame cut off.
• Reframe goes further, using AI to shift the physical perspective of an already-captured shot.
• Natural-language editing lets you ask the assistant to crop a picture, fix grammar in text, or write a custom shortcut by speaking plainly.

Each of those features depends on the same rented reasoning engine, which is why the camera upgrades and the assistant overhaul are really one story wearing two badges.

What iOS 27 Siri Means for Your iPhone

For most people, the practical question is simple: a Siri that finally does what Apple said it would two years ago, with a layout cribbed from the chatbots they already use. The trade-off sits in the plumbing rather than the screen.

Here is the shape of the rollout in numbers:

• June 8 is the WWDC keynote where the redesign gets its official reveal.
• September is the target window for the consumer release, alongside the iPhone 18 Pro lineup and Apple’s first foldable.
• 15 years is roughly how long Siri has gone without an overhaul this deep.
• The early intelligence upgrades already arrived in stages, part of a two-step plan covered in our breakdown of Apple’s staged Siri AI roadmap.

The cost is strategic rather than personal. Apple keeps the data inside its own walls, but it no longer controls the most valuable component of its flagship feature, and it pays a rival to keep it running. That dependency is fine while the partnership holds. It becomes a pressure point the first time Google raises the price or ships a sharper model to its own Pixel phones before Cupertino gets it.

If the September launch delivers the responsiveness the renderings imply, Apple buys itself time to rebuild its own models behind the scenes while customers stop complaining. If it stumbles again, the company will have spent a billion dollars a year to look like it is following the field it once expected to lead.

Frequently Asked Questions

When does the redesigned Siri launch?

Apple will reveal the new Siri at its WWDC keynote on June 8, with a general consumer release targeted for September alongside the iPhone 18 Pro and iPhone 18 Pro Max. Early intelligence features rolled out before that in staged updates.

Which AI model powers the new Siri?

A custom 1.2 trillion parameter version of Google’s Gemini handles the complex reasoning, built specifically for Siri and Apple Intelligence. It is roughly eight times larger than the cloud model Apple developed on its own.

Does Google see my Siri data?

No, based on Apple’s stated design. The licensed Gemini model runs inside Apple’s Private Cloud Compute servers rather than on Google Cloud, so queries are processed on Apple hardware and discarded afterward instead of being shared with Google.

What changes about how I open Siri?

Swiping down from the top center of the screen opens a new “Search or Ask” panel, and Notification Center moves to a swipe from the top-left corner. Siri animations and answer cards now appear inside the Dynamic Island, and a standalone app stores your conversation history.

What are the new camera tools?

Two editing features called Extend and Reframe use AI to generate scenery beyond a photo’s borders and to shift its perspective, while a new visual mode replaces Visual Intelligence and lets you run images through reverse search or third-party AI agents.