Apple confirmed a multi-year deal with Google to power a new Siri with Gemini. The rollout started in iOS 26.4 with the full overhaul arriving in September with iOS 27 and iPhone 18.
Siri was a voice command interface. Gemini-powered Siri is a decision-making layer. Those are different products with different implications for discovery.
Apple is paying Google approximately $1 billion a year to put Gemini inside Siri. That's the financial headline. The strategic implications are larger.
The deal was confirmed in January 2026. iOS 26.4, released this spring, already uses Gemini for improved context awareness and on-screen recognition. The full overhaul, described by Apple as Full Conversational Siri, arrives with iOS 27 and the iPhone 18 in September.
The features Apple is building are a direct response to losing ground to ChatGPT and Perplexity in conversational search. Persistent chat history, deeper cross-app awareness, the ability to chain multiple commands into a single request, and a standalone Siri app that behaves more like a chatbot than a voice assistant. Processing runs in Private Cloud Compute to maintain Apple's privacy commitments.
Apple's approximate annual payment to Google for the Gemini integration licence powering the new Siri
For marketers, the question isn't about the technology. It's about what happens to intent when a billion-plus Apple users have a genuinely capable AI assistant on their device.
The shift matters because iPhone users currently account for a disproportionate share of high-value consumer intent in Australia and globally. If those users start using Siri to research products, compare options and make decisions, the discovery layer moves from Google Search results to Gemini's model responses. That's Google's model, filtered through Apple's product experience and privacy architecture.
The competitive dynamic is genuinely unusual. Google is being paid $1 billion a year to power a product that may reduce Google Search queries on Apple devices. Apple is using a competitor's model to close the gap with ChatGPT. Neither company had a better option.
For marketers, the practical implication is the same as every AI search development before it: if your content isn't in the retrieval layer, you won't be in the answer.