WWDC 2026 Keynote

The Apple Worldwide Developers Conference (WWDC) 2026 keynote took place on Monday, June 8, 2026. This event marked a major transition point for Apple, serving as the final keynote for CEO Tim Cook before he steps down in August 2026, with John Ternus set to succeed him.

The presentation was strictly focused on software, user interface refinements, and the integration of artificial intelligence, with no hardware announcements.

---

🤖 Siri AI & Apple Intelligence

Siri AI (Powered by Google Gemini)

Apple introduced a ground-up rebuild of Siri, branded as Siri AI, backed by a strategic multi-year partnership with Google.

• Engine Partnership: Apple’s system-level assistant (Siri AI) and default cloud-based Apple Foundation Models are powered by Google’s Gemini models (referenced in Google IO 2026). Apple runs distilled versions of these models locally on-device and escalates complex reasoning tasks securely via its own Private Cloud Compute (PCC).
• Privacy Protections: Unlike normal web chatbots, user queries sent through Siri AI are processed with strict privacy guardrails: data is not stored long-term, not shared for ad targeting, and not used to train Google’s public models.
• Screen Awareness & Context: Siri can see and understand content on the user’s screen contextually, maintain multi-turn history, retrieve personal data (e.g., info from emails, calendar, messages), and execute multi-app workflows.
• AI Extensions System: In iOS 27 and macOS 27, users can choose alternative default chatbots (like Claude, ChatGPT, or the direct Gemini chat interface) to handle writing assistance and dynamic image creation tasks.
• Chatbot Interface: Can be summoned via the power button, the Dynamic Island, or through a dedicated new Siri app interface.
• Timeline & Launch: Developer betas are available now (June 8, 2026). The public beta is scheduled to launch in Fall 2026 (expected September 2026 alongside iOS 27). Siri AI will initially bypass the EU and China due to regulatory barriers.
• Pricing & Subscription: Free. Siri AI and Apple Intelligence features are included as free system updates. However, certain advanced cloud-processing limits may scale with iCloud+ subscription tiers, and third-party developer integrations (like using Claude in custom apps) may require separate developer API keys or accounts.
• Device Requirements:
  • Standard Siri AI & Apple Intelligence: Supported on iPhone 15 Pro, iPhone 15 Pro Max, or any iPhone 16 / iPhone 17 (standard models require at least 8GB RAM).
  • Advanced Features (12GB RAM limit): Requires iPhone 17 Pro, iPhone 17 Pro Max, iPhone Air, or newer to run Apple’s most powerful on-device model locally. Base models like the standard iPhone 17 (8GB RAM) fallback to Private Cloud Compute for these specific features:
    • Expressive Siri Voices: Real-time generation of natural, customized voices with dynamic pace adjustments.
    • Advanced Dictation: Next-generation offline dictation with ultra-high accuracy and automatic punctuation.
    • Largest On-Device LLM: Apple’s top-tier reasoning model for deep context understanding without cloud latency.
  • iPads & Macs: Requires an M1 chip or newer (iPad mini requires A17 Pro or newer).

Apple Intelligence Upgrades

• Photo Editing:
  • Clean Up: Erases background distractions using spatial models.
  • Extend & Reframe: Uses AI to generate background elements beyond image borders or adjust camera angles post-capture.
• Smart Notifications: AI-generated summaries and priority alerts integrated deeply into system apps like Home and Messages.

---

💻 Operating System Releases

iOS 27

• Year-Based Naming: Follows Apple’s transition to year-based numbering (matching the calendar year the release runs through).
• Liquid Glass Refinements: Offers a transparency/opacity slider for the UI based on user feedback from previous versions.
• Compatibility: Supported on iPhone 11 and newer devices.

macOS 27 (“Golden Gate”)

• Siri AI Support: Fully supports Siri AI and Apple Intelligence.
• Spotlight & App Integration: AI queries can be typed directly into Spotlight and are automatically routed to Siri AI. A dedicated Siri app is introduced for managing cross-device chat history via iCloud.
• Productivity & File Focus: Optimizes Siri AI for desktop workflows, enabling deep document editing, Finder-based file organization, and desktop-level task hand-offs.
• Apple Silicon Exclusive: Drops support for all Intel-based Macs. It runs exclusively on Apple Silicon (M-series and newer).
• System Requirements: General Apple Intelligence features require at least M1 with 8GB RAM. Advanced on-device features (expressive voice, advanced dictation) require M3/M4 and at least 12GB of unified memory.
• Finder Rebuild: Features a rearchitected content indexing engine to resolve search and stability issues.

iPadOS 27

• UI Customization: Users can lock the menu bar on-screen, resize iPhone apps, and see active app names in the status bar.
• Safari Tab Topics: Auto-clusters browser tabs into logical subject categories using AI.
• System Requirements: Advanced Apple Intelligence features require an M4 chip or newer and at least 12GB of RAM.

watchOS 27 & visionOS 27

• watchOS 27: Drops support for Series 8, Ultra 1, and SE 2. Integrates Siri AI and health monitoring refinements.
• visionOS 27: Introduces curved-window layout controls, eye-tracking interactive expansions for notifications, and converts panorama photos into immersive 3D spatial environments.

---

🛠️ Developer Tools & Frameworks

Craig Federighi highlighted WWDC 2026 as the launchpad for Agentic Coding on Apple platforms, featuring deep integrations with cloud-based LLM providers, most notably Anthropic’s Claude:

🧡 Anthropic Claude & Apple Ecosystem Integration

The integration of Anthropic’s Claude into Apple’s ecosystem was a major highlight for developer tooling and APIs:

1. Hybrid App Architecture via the Foundation Models Framework

Apple expanded the Foundation Models framework into a unified Swift API that natively supports both on-device models and cloud-based models:

• Language Model Protocol: Developers can now interact with cloud models like Claude and Gemini using the same native Swift API as Apple’s local on-device models by conforming to the LanguageModel protocol.
• Task Hand-off: Apps can leverage Apple’s local models for low-latency, privacy-sensitive actions (like text extraction or local classification) and seamlessly hand off complex reasoning, code generation, or web-connected research tasks to external cloud engines.
• Provider Libraries & Packages:
  • Anthropic (Claude): Released an official Swift package to integrate Claude with the framework, automating connection management, API keys, streaming, and tool calls.
  • Google (Gemini): Released direct integration support via the Firebase Apple SDK, allowing Gemini models to plug directly into the Swift framework.
  • OpenAI (ChatGPT): Has not released an official first-party package conforming to Apple’s native LanguageModel protocol yet (developers still use direct HTTP calls or community client libraries).
• Type-Safe Structured Output: By using the framework’s @Generable annotations, developers can receive structured Swift types directly from Claude, facilitating safe data exchange between on-device processing and Claude API calls.

2. Native Xcode Agentic Coding (Xcode 27)

Building on the native integration of the Claude Agent SDK in Xcode 26.3 earlier in the year, Apple announced Xcode 27 with advanced multi-model support:

• Multi-Model Agent Workflow: Xcode 27 brings coding agents from major providers (Claude, Google Gemini, OpenAI) directly into the IDE workflow. Developers can select their preferred model.
• Autonomous Capabilities: The Claude Agent in Xcode can reason across entire workspace directories, plan architecture, and edit multiple files.
• Tool Integration & MCP: Using the Model Context Protocol (MCP), the agent can execute local commands, run tests, and automatically capture Xcode Previews to visually verify layout designs (particularly in SwiftUI).
• Interactive Canvas: A side-by-side canvas presents markdown plans, diffs, and live simulator previews.
• Onboarding via CLAUDE.md: Xcode agents read a CLAUDE.md file in the root of the project to dynamically adapt to the project’s specific styling, state management, and architecture guidelines.

⚙️ Framework & Tooling Upgrades

• Core AI Framework: A new framework optimizing on-device model execution speed and efficiency on Apple silicon (M-series / A-series chips).
• Simulator Automation: Coding assistants can programmatically control simulated devices, run UI tests, and inspect state hierarchies.
• Expanded App Intents: Deeper integration hooks for Siri AI to invoke multi-step operations within third-party apps.