WWDC 2026 wrapped up its keynote on June 8, and this one felt different. It was Tim Cook’s last keynote as CEO before he hands the reins to John Ternus on September 1. But the bigger story for those of us who write code? Apple finally went all-in on AI — and most of the news was aimed squarely at developers.
If WWDC 2025 was about the “Liquid Glass” design language, WWDC 2026 was about putting real, on-device intelligence into your hands through native Swift APIs. Let’s walk through everything that matters, starting with the parts you’ll actually be shipping.
Quick note before we dive in: developer betas are available now (June 8), public betas land next month (July), and the public release ships this fall. All the usual version bumps are here — iOS 27, iPadOS 27, macOS 27 (codenamed Golden Gate), watchOS 27, tvOS 27, and visionOS 27.
1. Foundation Models Framework Grows Up
The Foundation Models framework first arrived at WWDC 2025 with a beautifully simple promise: three lines of Swift to call a ~3B parameter on-device model — privately, offline, with no API key and no billing. This year, Apple made it genuinely powerful.
What’s new:
- Vision inputs. You can now pass images as input alongside text. That opens the door to multimodal features without reaching for a cloud API.
- Custom AI skills. You can define your own skills and plug them into the model.
- Unified Swift API for on-device and server models. The same API surface now lets you tap both the local model and Apple’s Private Cloud Compute, so you choose where inference happens.
- A bigger, smarter base model. The new on-device model is distilled from Google’s Gemini family and is reportedly significantly larger than the 2025 version, with an expanded context window.
- On-device fine-tuning. You can now fine-tune privately, with no data leaving the device.
The best part: the original three-line Swift API hasn’t changed. Structured outputs, streaming, tool calling, and session management all live in the same surface — and it remains free for developers. No API key, no rate limits, no bill.
Dynamic Profiles: the standout new API
The most interesting addition is Dynamic Profiles — a set of new declarative APIs for building adaptive AI experiences with far less code. Instead of locking a session to a fixed model, set of tools, and instructions, you can now orchestrate skills and subagents, swap tools in and out, and update instructions on the fly during a live session.
In plain terms: your AI session can now evolve as the user’s task evolves, without you tearing it down and rebuilding it. If you’re building anything agent-like, this is the API to learn first.
2. Core AI Replaces Core ML
This is the headline shift, and it’s a big one: after nine years, Core ML is being replaced by Core AI.
Core ML was built back in 2017 for image classifiers, text regressors, and tree ensembles running on the Neural Engine. It did that job well for a decade — but it was never designed for large language models, streaming token generation, or agent-style tool calling, which is exactly what we’re all building now.
What Core AI gives you:
- The ability to run third-party AI models locally on Apple devices, taking full advantage of Apple Silicon’s performance while keeping data on-device.
- Native support for the Model Context Protocol (MCP).
- Foundation Models becomes the developer-facing API surface within Core AI.
The migration reality check: Core ML doesn’t connect to any of the new APIs. If you’re adding generative features to an existing app, you’ll be writing Core AI code from scratch. If you have .mlmodel files in production, they'll keep working for now — but new AI work should start on Core AI. Plan your migration accordingly.
3. Connect Your App to Siri AI via App Intents
Here’s a distribution opportunity that’s easy to overlook. Through App Intents, you can now wire your app directly into the new Siri AI, so users can find information and perform actions inside your app simply by asking Siri.
With two billion-plus Apple devices expected to be running an AI-capable Siri, App Intents effectively becomes a new discovery and engagement channel. If your app exposes the right intents, it becomes part of how people get things done — without ever opening it manually. Worth prioritizing.
There’s also a new Extensions framework that lets users pick their preferred AI provider (Claude, ChatGPT, or Gemini are supported at launch). If your product is an AI assistant or model, the Extension SDK is open in the developer beta today.
4. Xcode 27 Goes Fully Agentic
Xcode quietly became the most AI-forward IDE Apple has ever shipped. The move is from simple autocomplete to agentic coding — the IDE can now take on entire tasks, not just suggest the next line.
The architecture is dual-engine:
- A local Neural Engine model handles real-time Swift suggestions on-device.
- A cloud routing layer handles heavier analysis, routing to Anthropic’s Claude, Google’s Gemini, or OpenAI’s agents — your choice.
What the coding agent can actually do:
- Simulate entire apps and inspect visual changes through live previews.
- Write and run tests.
- Interact with Playgrounds.
- Operate the iOS Simulator through a new Device Hub.
- Localize an entire app automatically.
- Handle multi-turn conversations with streaming responses, plus a canvas that renders Markdown alongside your code changes and previews.
You onboard through Xcode’s new Intelligence settings panel, and the agent ships with 20+ MCP tools already wired up. Notably, this means any MCP-compatible agent — including Claude Code and OpenAI Codex — can invoke Xcode’s build, test, and diagnostics operations. Apple also folded in “vibe coding” support via Gemini for those who like to build conversationally.
5. Siri AI: The Brain Transplant
The user-facing centerpiece is the rebuilt Siri AI, now powered under the hood by Google Gemini (reportedly a custom large model trained for Apple). The new Siri is more capable, far more conversational, and gains visual intelligence — and it lives in a standalone app in addition to working across your existing apps.
Apple was emphatic about privacy here. Craig Federighi framed privacy in AI as non-negotiable, stressing that your data is used only to fulfill your request and that outside experts can continue to verify that claim.
One critical caveat for developers in certain regions: Siri AI will not ship in the European Union or China with iOS 27 or iPadOS 27 at launch. Apple is pointing to the EU’s Digital Markets Act as the blocker. The good news for builders: the Foundation Models developer API is not subject to that geographic restriction — you can build and test against it globally. And Siri AI will still be available in the EU on macOS 27, visionOS 27, and watchOS 27.
6. iOS 27 Reaches All the Way Back to the iPhone 11
Apple called iOS 27 the most widely available iOS release ever — it supports every device from the iPhone 11 onward. For developers, that’s a huge addressable base, but it also means your AI-heavy features need graceful fallbacks: Apple Intelligence and Siri AI themselves require newer silicon (iPhone 16 and later, iPhone 15 Pro/Pro Max, M1+ iPads and Macs, and so on), even though the OS runs on older hardware.
Test your intelligence features against the capability matrix, not just the OS version.
7. Performance Wins You Get for Free
A lot of this year’s speed improvements come at the system level — meaning your app benefits without any code changes:
- App launches up to 30% faster.
- New photos appear in the camera roll up to 70% faster.
- AirDrop transfers up to 80% faster, with quicker Mail loading and Apple Music playback too.
- A modified CPU scheduler makes older iPhones (think iPhone 11 and up) feel noticeably snappier.
- Search was rebuilt from the foundation up — the infrastructure powering Spotlight, Mail, and Photos is now more stable and efficient, indexing new files and data almost immediately.
- iCloud shared albums now support full-resolution photos, including on Android and Windows.
8. Liquid Glass: Now With a Volume Knob
If you (or your users) found last year’s Liquid Glass a bit much, relief is here. Apple isn’t ditching the aesthetic, but it’s adding an opacity slider so users can dial transparency up or down — or turn it off. The design is more readable and customizable by default.
On macOS Golden Gate, expect more uniform toolbars across apps, sidebars that stretch to the screen edges to reduce distraction, tighter window corner radii, and refreshed app icons using a new layered approach. If your app leans on the system design language, give it a pass to make sure it still looks right across the new defaults.
9. Trust, Safety, and Child Protection APIs
Apple introduced a set of new APIs to help protect children when they use your apps, including the ability to tailor experiences to a user’s age range. If your app serves a broad audience — or specifically younger users — these APIs are worth integrating early to stay ahead of both user expectations and regulatory pressure.
The Developer Action List
If you only have an afternoon this week, here’s where to spend it:
- Enable the Foundation Models framework and try the three-line on-device call if you haven’t yet. It’s free and shipping.
- Learn Dynamic Profiles — this is the API that makes adaptive, agent-style features practical.
- Audit your Core ML usage and plan a Core AI migration path for any generative features.
- Add App Intents to plug your app into Siri AI — it’s a real distribution channel now.
- Turn on Xcode 27’s agentic coding and connect the MCP tools.
- Check your capability matrix so AI features degrade gracefully on older, iOS 27-supported hardware.
- Grab the developer beta today, and watch for public betas next month ahead of the fall release.
Summary
For years, Apple’s on-device ML story never quite matched what the Python and cloud-API ecosystems offered. WWDC 2026 changes that. Between a grown-up Foundation Models framework, Core AI, system-wide MCP, and a genuinely agentic Xcode, Apple is now shipping tools that compete directly with what most of us have been reaching for cloud APIs to accomplish — with zero inference cost and a strong privacy story baked in.
It’s a fitting send-off for Tim Cook’s final keynote, and a clear signal of where the platform is heading. For iOS developers, the message is simple: the on-device AI era just became real, and the entry cost is three lines of Swift.
Time to build.
Level Up Your Mobile Developer Interview !
iOS Developer Interview Handbook
From Swift fundamentals to advanced iOS concepts — a complete handbook to help you prepare smartly and confidently.
👉 Explore the book:
https://medium.com/@anandgaur2207/crack-ios-developer-interviews-with-confidence-the-complete-ios-developer-handbook-f1eabc3d7a21
Cracking the Mobile System Design Interview Book
Your complete practical guide to mastering Mobile System Design Interviews — covering scalable architecture, Android & iOS system design concepts, high-level design strategies, low-level design patterns, performance optimization, offline-first architecture, real-world case.
👉 Grab your copy now:
https://medium.com/@anandgaur2207/cracking-the-mobile-system-design-interview-book-8ff043db0359
Mastering AI for Android Developers
Your complete hands-on guide to integrating AI into Android apps — covering Generative AI, LLMs, on-device intelligence, AI APIs, real-world use cases, and practical implementation with modern Android development.
👉 Grab your copy now:
https://medium.com/@anandgaur2207/mastering-ai-for-android-developers-5cc6d62e7d21
Crack Android Interviews Like a Pro
Your complete Android interview preparation book — packed with real questions, deep explanations, and practical insights to help you stand out.
👉 Grab your copy now:
https://medium.com/@anandgaur2207/crack-android-interviews-with-confidence-the-only-handbook-youll-need-b87ec525f19c
Flutter Developer Interview Handbook
Ace your next Flutter interview with scenario-based questions, detailed explanations, and hands-on examples that make you stand out.
👉 Explore the book:
https://medium.com/@anandgaur2207/crack-flutter-developer-interviews-with-confidence-the-complete-flutter-developer-interview-6cb53996832c
React Native Developer Interview Handbook
Crack your next React Native interview with confidence!
This guide is packed with scenario-based questions, detailed explanations, and hands-on examples to help you stand out and succeed.
👉 Explore the book:
https://medium.com/@anandgaur2207/react-native-interview-crack-your-next-interview-with-confidence-0d7255a20fe1
Need 1:1 Career Guidance or Mentorship?
If you’re looking for personalized guidance, interview preparation help, or just want to talk about your career path in mobile development — you can book a 1:1 session with me on Topmate.
I’ve helped many developers grow in their careers, switch jobs, and gain clarity with focused mentorship. Looking forward to helping you too!
Found this helpful? Don’t forgot to clap 👏 and follow me for more such useful articles about Android development and Kotlin or buy us a coffee here ☕
If you need any help related to Mobile app development. I’m always happy to help you.
Follow me on: