Why Apple may put Google Gemini at the center of Siri

Apple is reportedly close to a licensing deal with Google that would bring Gemini into core Siri functions. The plan would use Google’s model on Apple servers, while Apple continues work on its own AI models for the longer term.

WTF Index IDIOCRACY
◄ Terminator 1 Idiocracy 2 ►

The story mainly points to deeper reliance on AI assistants for summarizing and planning, with little evidence of major danger or loss of control.

Why Apple may put Google Gemini at the center of Siri

Apple is reportedly preparing a major change to Siri: a licensing deal with Google that would put Gemini behind some of the assistant’s most important new capabilities. According to Bloomberg, the agreement is valued at about $1 billion per year and would give Apple access to Gemini for core Siri functions.

The move shows how urgently Apple is trying to improve its voice assistant after internal setbacks caused it to miss the initial wave of generative AI adoption. It also suggests a careful compromise: use Google’s model where it can help now, but keep the experience running inside Apple’s own infrastructure.

What Google Gemini would do for Siri

The reported deal centers on new "summarizer and planner" features. In practical terms, those functions are meant to help Siri collect, condense, and organize complex information more efficiently.

That matters because a modern AI assistant is expected to do more than answer short commands. It needs to handle messy requests, connect pieces of information, and turn them into something useful. The source article describes this as a central part of the plan for the new Siri.

The Gemini model Apple intends to use reportedly has around 1.2 trillion parameters. That would make it far larger than the 150-billion-parameter cloud AI Apple currently employs.

Parameter count is not the only measure of quality, but within the facts reported here, the contrast is important. Apple appears to be looking outside its own stack for a model powerful enough to support the new Siri experience it wants to ship.

The new Siri timeline

The revamped Siri is internally known as "Glenwood" and marketed under the codename "Linwood." It is scheduled to debut with iOS 26.4 in spring 2026.

The project is led by Mike Rockwell, who also heads Vision Pro development, and Craig Federighi, Apple’s software chief. The source also notes that some Siri components will continue to rely on Apple’s own AI technology.

That split is important. Based on the report, Apple is not simply replacing Siri with Gemini. Instead, it is expected to use Gemini for specific core functions while keeping other parts of the assistant tied to Apple’s own systems.

The result would be a hybrid approach. Google would provide a major language model capability, while Apple would still shape the product, the user experience, and at least part of the underlying AI stack.

How Apple plans to keep Google in the background

Even though Google is providing the model, the integration is reportedly designed to run on Apple’s servers. According to Bloomberg, all user data will remain entirely separate from Google’s infrastructure.

Apple also does not plan to publicly market the Google name. That is a notable contrast with the long-standing deal that makes Google the default search engine in Safari.

For users, the intended effect is simple: Siri may become more capable, but the assistant would still look and feel like an Apple product. The Google role would sit behind the scenes rather than becoming a visible part of the Siri brand.

The current agreement also does not include Google’s search service. Earlier talks about embedding Gemini as a full chatbot failed, and the reported deal is focused on using the model for Siri’s core functions instead.

Why Apple chose Google, and what comes next

Before settling on Google, Apple reportedly explored partnerships with OpenAI (ChatGPT) and Anthropic (Claude). It ultimately chose Gemini.

The source does not say that Apple is abandoning its own AI ambitions. In fact, the long-term plan points in the opposite direction. Apple intends to rely on its own AI models again.

A 1-trillion-parameter model is already in development and could be ready for consumer devices as soon as next year. Apple hopes that system will eventually reach Gemini-level quality.

There is still risk in that plan. Apple has lost several key AI researchers to Meta in recent months, which could slow progress. That makes the Google deal look less like a final destination and more like a bridge while Apple continues building.

China may require a different Siri strategy

The reported Gemini plan may not apply everywhere. In China, Apple will reportedly take a different approach.

Gemini is unlikely to be used there. Instead, Apple plans to deploy its in-house models filtered through an additional Alibaba mechanism required by Chinese regulators.

The company is also considering a separate partnership with local AI firm Baidu. That means Siri’s AI foundation could vary by market, with Gemini powering the assistant in some places while other systems support it elsewhere.

Taken together, the report describes a pragmatic shift for Apple. Siri needs stronger generative AI capabilities, Google has a model Apple reportedly wants to license, and Apple still wants to keep control over infrastructure, branding, and its longer-term model roadmap.