Apple announced a partnership with Google to rebuild Siri using Google’s 1.2 trillion parameter Gemini model, running on Apple’s Private Cloud Compute infrastructure for privacy. The deal, reported by CNBC in January 2026, is valued at up to $5 billion with Apple paying approximately $1 billion annually for Gemini access.
What Apple Admitted
By partnering with its biggest rival for the core intelligence behind Siri, Apple acknowledged what users had complained about for years: its own AI was not competitive. The new Siri gains complex reasoning, multi-step planning, screen awareness, and the ability to take actions within apps — capabilities that Google’s Gemini and OpenAI’s GPT models have demonstrated but Apple’s on-device models could not match.
The privacy architecture is the key compromise. Gemini operates as a “white-labeled” engine with no Google branding visible anywhere in the Siri interface. The model runs on Apple’s Private Cloud Compute servers, meaning user data remains isolated from Google’s infrastructure. Apple is renting Google’s intelligence while keeping it locked in Apple’s house — a deal structure that preserves Apple’s privacy narrative while conceding its AI capability limitations.
The Delay Problem
The Gemini-powered Siri upgrade did not ship with iOS 26.4 when it launched March 23, 2026, despite the announcement 10 weeks earlier. 9to5Mac reported the features were pushed to iOS 26.5 (May) and iOS 27 (September) due to quality issues. The delay reinforces the concern: even with Google’s most capable model, integrating conversational AI into a mobile operating system at Apple’s quality bar is proving harder than either company expected.
Samsung, meanwhile, is targeting 800 million devices with Gemini-backed AI features by end of 2026 — doubling from roughly 400 million, as confirmed by Samsung co-CEO TM Roh at CES 2026. Google’s model now powers or will power the assistants on both major mobile platforms, giving it a distribution advantage that no other AI company can match.
