Newsletter Subscribe
Enter your email address below and subscribe to our newsletter
Enter your email address below and subscribe to our newsletter

Apple used its WWDC 2026 keynote on June 8 to introduce Siri AI, a ground-up rebuild of the voice assistant that the company said was developed in collaboration with Google using the Gemini family of models. The announcement marks Apple’s most ambitious attempt to close the gap with rivals in the AI assistant space after years of criticism that Siri lagged behind ChatGPT and other competitors.
Rather than bolting new capabilities onto the old Siri, Apple built what it calls a five-tier system of Apple Foundation Models ranging from lightweight on-device processing to a cloud model derived from Gemini Pro with approximately one trillion parameters, according to PCMag. Craig Federighi, Apple’s senior vice president of software engineering, told attendees the company “embarked on a deep collaboration with Google” to develop the next generation of models.pcmag
Despite the Google partnership, Apple emphasized that the end product runs on its own infrastructure. Amar Subanya, Apple’s vice president of AI engineering and a former Google Gemini team lead, said Siri AI was trained using reinforcement learning and refined using outputs from Gemini’s most advanced models, but does not use Google’s deployment framework or search index. Most requests are processed on-device, with heavier queries routed through Apple’s Private Cloud Compute architecture running on Nvidia GPUs in Google Cloud.letsdatascience
The revamped assistant supports conversational continuity, on-screen awareness, and multi-step reasoning across iPhone, iPad, and Mac. A dedicated Siri app will let users revisit conversation history, and the assistant can surface information from messages, emails, and photos without users needing to specify where to look.engadget
Days after the keynote, The Next Web reported that the iOS 27 developer beta contains a dormant Extensions framework allowing users to swap between ChatGPT, Claude, and Gemini inside Siri. Bloomberg’s Mark Gurman reported that Apple has held entitlement discussions with OpenAI, Anthropic, and Google, though the feature was not announced on stage. The framework includes a settings panel and a dedicated App Store section, both currently disabled on Apple’s backend.thenextweb
The omission from the keynote appears tied to multiple headwinds: Apple confirmed Siri AI will not ship in the EU on iPhone or iPad due to Digital Markets Act concerns, and Bloomberg reported that OpenAI is exploring legal options over its existing Siri partnership.letsdatascience
Siri AI is available now in the iOS 27 developer beta and is expected to reach a public beta later this summer, with a stable release this fall alongside new hardware. The assistant will initially support English only and will not be available in the EU on iOS or iPadOS, nor in China, though EU users will have access on macOS and visionOS. Apple’s most advanced on-device voice features require an iPhone 16 or later, or recent iPad and Mac models with sufficient unified memory.cnbc