Apple WWDC 2026: Will Cupertino Finally Deliver an AI Siri Worthy of the Name?
Today at 7:00 PM, Apple kicks off its WWDC 2026 developer conference, and it is arguably the most important keynote appearance in years. After the chaotic launch of “Apple Intelligence” in 2024, the company has to prove it can keep pace in the AI race with OpenAI, Google, and Anthropic. At the center of it all is a completely rebuilt Siri. Because what has been delivered so far is not what users expected; Apple even paid $250 million in a settlement over unfulfilled Siri promises.
Two years after the big promises around a “context-aware” Siri that was never shipped, this year’s developer conference is therefore being seen as Apple’s AI liberation strike. According to consistent reports, the keynote will be less a hardware event than a software restart—with Siri as the lead act. Here are the most important expected announcements in detail.
1. The new Siri runs on Google Gemini
The most consequential deal of all: for the Siri overhaul, Apple is licensing a custom-built Gemini model from Google. Bloomberg reporter Mark Gurman reports that Apple is paying around $1 billion per year for this—for a customized model with roughly 1.2 trillion parameters that powers Siri’s cloud functions. The partnership had already been announced in January 2026. Importantly: the Gemini queries are said to run through Apple’s own Private Cloud Compute infrastructure, so that Google gains no access to user data.
2. Siri becomes a standalone app
For the first time, Siri is set to get a real app “shell”—similar to ChatGPT, Gemini, or Claude. Gurman published advance renderings showing a standalone Siri app on iPhone, iPad, and Mac: a dark interface, a text input field, a microphone toggle for voice mode, and a paperclip icon for attaching images and files. Conversations are saved as a history (with an optional expiration date) and synced via iCloud. On top of that comes a new system-wide “Search or Ask” gesture, which opens an AI search bar out of the Dynamic Island via a swipe from the top.
3. “Extensions”: Users can choose the AI model themselves
Apple is opening Siri up to third-party models. Through a feature called Extensions, it should in the future be possible to integrate not only the existing ChatGPT but also Google Gemini and Anthropic’s Claude—users can then select which model answers their queries. Notably: Claude and ChatGPT lost out in the model selection for the deeper Siri integration, but come back into the system “through the back door” via Extensions.
4. Siri as a system-wide agent
The new Siri is supposed to do considerably more than set timers. According to Gurman, it will draw on personal data—emails, calendar, contacts, notes, photos—”understand” on-screen content (On-Screen Awareness), and carry out multi-step tasks across several apps. For example: drafting an email based on past messages, or merging information from the web and one’s own documents. The new “Ask Siri” feature is meant to turn the assistant into a system-wide AI agent.
5. AI agents in the App Store
Perhaps even more significant structurally than Siri: according to The Information, Apple is working on a set of rules that would permit AI agent apps in the App Store—that is, software that completes tasks independently, such as making reservations, managing to-dos, editing documents, or controlling smart home devices. The background: Apple had blocked “vibe-coding” apps in March for violating Store guidelines and is now seeking a middle path that preserves security and data protection.
6. Camera and Photos get AI upgrades
Also expected is a new Visual Intelligence mode directly in the Camera app—as its own mode alongside Photo, Video, and Panorama. With it, you can point at an object to ask Siri what it is or who could repair it; the previous Visual Intelligence function via the Camera Control button is being replaced. In the Photos app, new AI tools such as “Reframe” and “Extend” are set to be added. There are also reports of more realistic image generation in Image Playground, new Genmoji designs, and AI-generated wallpapers.
7. Natural language for shortcuts
A quieter but practical innovation: in iOS 27, Shortcuts can reportedly be described in plain text, instead of laboriously clicking them together from building blocks. This could make one of the iPhone’s most underrated productivity features accessible to a broader audience.
8. Barely any new hardware—and Tim Cook’s last keynote
New devices are considered unlikely. The rumored HomePad (a HomePod with a display) is said to be finished but is being held back because it depends heavily on the new Siri. A Mac mini M5 or an iPhone Fold are also unlikely to appear on stage. In any case, it will be Tim Cook’s last major keynote as CEO—he is set to step down on September 1, with hardware chief John Ternus expected to take over.
Broad rollout expected for the fall
Apple is going into this WWDC with its back against the wall: only in May did the company agree to a $250 million settlement with iPhone buyers who felt deceived by the never-delivered Siri promises from the iPhone 16 advertising. The central question this evening is therefore not whether Apple will show a new Siri, but whether it will actually ship this time—and how much of its own AI narrative Apple is handing over to Google. The public beta is expected in July, with the broad rollout alongside iOS 27 in the fall.

