Apple Unveils Gemini-Powered Siri and iOS 27 at WWDC 2026

Bloomberg reports that Apple will use its WWDC 2026 keynote on Monday, June 8 to introduce a refreshed artificial intelligence strategy and major updates to its operating systems, collectively labeled iOS 27, iPadOS 27, macOS 27, tvOS 27, watchOS 27, and visionOS 27 (Bloomberg). TechCrunch and Bloomberg report leaked renders and coverage showing a rebuilt, standalone Siri app and an updated assistant experience that surfaces from the Dynamic Island and a new system-wide "Search or Ask" interface (TechCrunch; Bloomberg). Multiple outlets, including MacRumors and AppleInsider, report that the revamped Siri will be able to access personal data such as emails, messages, files, and photos to complete tasks, will have on-screen awareness to act on visible content, and will perform deeper actions across apps (MacRumors; AppleInsider). Apple has not publicly confirmed these features ahead of the keynote.

Bloomberg's Mark Gurman reports that Apple will pay Google roughly $1 billion per year for a custom, roughly 1.2-trillion-parameter Gemini model to run Siri's cloud-based capabilities, while Apple continues to develop its own on-device models for privacy-sensitive tasks (Bloomberg). TechCrunch frames the approach as a hybrid that pairs cloud LLM capability with local processing (TechCrunch). 9to5Mac, citing Gurman, describes a new gesture that opens a system-wide "Search or Ask" bar with a microphone toggle and the option to switch between Siri and third-party models such as ChatGPT or Gemini for query handling (9to5Mac). Bloomberg and Tom's Guide report the new Siri is expected to expose a chatbot-like conversation history, uploads, and formatted card-style results in the UI (Bloomberg; Tom's Guide).

Editorial analysis: companies that pair an external LLM with on-device models typically pursue a hybrid architecture to balance raw capability, latency, and privacy. The reported features, personal-context access, on-screen awareness, and an uploads-enabled chat interface, mirror patterns already seen in competing assistant products that combine user-data plumbing with multimodal LLMs. A built-in toggle between a default vendor model and third-party models also echoes prior platform-level search arrangements, signaling a design that treats the underlying model as an interchangeable backend service rather than a single proprietary stack.

For practitioners, the developer-facing questions that will matter once the keynote and betas arrive include:

• •The scope of developer APIs and entitlements for personal-context access, as reported by MacRumors and AppleInsider.
• •The technical boundary between on-device models and cloud-served Gemini calls, including latency and integration points described in TechCrunch reporting.
• •Privacy and data-flow details for uploads, chat history, and cross-app actions, which multiple outlets list as central to the new assistant (MacRumors; Bloomberg).
• •Beta timing and compatibility: AppleInsider reports developer beta builds typically follow the keynote, letting teams test integrations and measure on-device resource use.

Editorial analysis: observers should treat the pre-keynote reports as expectations rather than confirmed specifications. The actual developer experience and platform constraints will only be clear once Apple presents at WWDC and ships betas and API documentation.