Following the long-awaited overhaul, the company says the system will bring deeper intelligence across its devices, but it will not launch everywhere at the same time.

The update focuses on a more conversational assistant that can understand context across apps and screens. It also allows users to interact with Siri in a more continuous way, including through a dedicated interface that stores recent interactions privately.

Apple says Siri AI can now search across Messages, Mail, Photos and Calendar to surface personal information when needed. It also reads what is on a device screen and responds based on that content. The assistant is also designed to work across iPhone, iPad, Mac, Apple Watch, AirPods and Vision Pro.

A camera-linked feature adds another layer, where users can point their device at objects and ask Siri to interpret them. Apple says this can include tasks such as analysing food items or helping split shared bills.

At the core of the system is what Apple calls “Apple Intelligence”, built on its own Foundation Models. The company says some tasks will run on-device, while more complex requests will use Private Cloud Compute.

Apple stated that “Privacy at Every Step,” is still very much central to the design, explaining that external support from Google’s Gemini model helps with some cloud-based functions. However, the company maintains that user data is not stored in a way that links back to individuals.

Despite the rollout, access will be restricted in key markets at launch. Users in the European Union will not receive Siri AI on iPhone or iPad when iOS 27 and related updates arrive later this year. China will also miss the initial rollout.

Apple links the European restriction to regulatory demands under the Digital Markets Act. The company argues that it cannot safely integrate the system under current conditions.

In its statement, it said regulators require “direct access to users’ private data – and the ability to directly control other installed applications – as soon as Siri AI is made available in the EU without the necessary safeguards to ensure user and data safety”.

The company also referenced its earlier proposal to regulators, describing a “Trusted System Agent” approach, which it says was not accepted. Apple maintains that the rejected model would have allowed safer integration with competing assistants.

Craig Federighi, Apple’s senior vice president of software engineering, said EU users will not get Siri AI on iPhone or iPad with the new operating system releases this year. He confirmed the restriction during the announcement window.

The company also stated that “extreme interpretation of the DMA” influenced its decision to hold back the feature. It added that it sees “clear dangers to EU users” under the current framework.

Even so, Apple confirmed that EU users will still access Siri AI on Mac computers and Vision Pro headsets. Those platforms will run macOS 27 and visionOS 27 with the feature included.

In China, Apple says approvals are still in progress, and no timeline has been confirmed for a full rollout. The company has not announced Siri AI availability for mainland China at launch.

It added that Hong Kong may receive earlier access, depending on language support and regulatory clearance. English-language beta testing is expected to begin in select regions later.

Apple’s strategy places Siri AI at the centre of its software ecosystem. The assistant will be integrated across iOS 27, macOS 27, watchOS 27, visionOS 27 and tvOS 27. The company is positioning this as a shift towards a more AI-led operating system design.

Developer access is already open through beta releases, with public testing expected around July 2026. A full release is expected later in the year, likely alongside the next iPhone cycle.

The agreement, announced on Monday, hands Google a function inside Apple’s ecosystem at a moment when Apple has had issues trying to scale in advanced software development. 

Under the deal, Google Gemini models will power the upcoming version of Siri and extend into other features tied to Apple Intelligence, a clear transition away from Apple’s tradition of relying almost entirely on its own tools.

Gemini already underpins key features in Samsung’s Galaxy devices, but Apple’s reach is much larger. With more than two billion active devices worldwide, Apple offers Google access to a scale that few platforms can match.

“After careful evaluation, Apple determined Google’s AI technology provides the most capable foundation for Apple Foundation Models,” Google said, adding that Gemini will support future Apple Intelligence features as well.

While neither company disclosed financial terms, industry reports put the deal at roughly $1 billion per year. If accurate, it would rank among the largest licensing agreements of its kind and underline how urgent Apple’s position has become. 

Gemini 3, Google’s latest flagship model, reportedly runs on about 1.2 trillion parameters. Apple’s internal models are believed to be far smaller, a gap that helps explain why Apple looked outside after repeated delays.

Those delays have been expensive. Since mid-2024, Apple has pushed back major improvements to Siri several times. Executive reshuffles followed, and the first wave of Apple’s generative tools failed to impress users or developers. 

Apple had already opened the door to outside help late last year by integrating ChatGPT into its devices. That arrangement allowed Siri to hand off complex questions to the chatbot, but only if users opted in. 

The new setup changes the balance. Gemini will sit much closer to the core of Apple’s system, while ChatGPT remains a secondary option.

“Apple’s decision to use Google’s Gemini models for Siri shifts OpenAI into a more supporting role, with ChatGPT remaining positioned for complex, opt-in queries rather than the default intelligence layer,” said Parth Talsania, CEO of Equisights Research.

However, Tesla chief executive Elon Musk wrote on X: “This seems like an unreasonable concentration of power for Google, given that (they) also have Android and Chrome.” Musk runs his own firm, xAI, which is investing heavily to compete with the biggest players.

Beyond the technology itself, the deal strengthens a commercial relationship that has lasted for years. Google already pays Apple tens of billions of dollars annually to remain the default search engine on iPhones and other devices.

Adding Gemini tightens that bond and makes Google even more embedded in Apple’s daily operations.

Investors welcomed the news. Alphabet’s market value climbed above $4 trillion on Monday, placing it in a small club alongside Nvidia, Microsoft and Apple itself. The company’s shares rose 65% last year, driven by growing trust in its strategy and speedy progress across text, image and video systems.

Google also moved quickly to address issues around data use. “Apple Intelligence will continue to run on Apple devices and Private Cloud Compute, while maintaining Apple’s industry-leading privacy standards,” the company said.

The agreement, according to Bloomberg, would be one of Apple’s biggest collaborations with an external technology partner in years.

For now, Apple plans to rely on Google’s large-scale model, which has 1.2 trillion parameters, to strengthen Siri’s processing power and decision-making. 

That’s nearly eight times more advanced than Apple’s current 150 billion-parameter cloud model. The company sees the deal as a temporary measure while it works to bring its own artificial intelligence system up to par.

Apple tested several models, including OpenAI’s ChatGPT and Anthropic’s Claude, before selecting Google’s Gemini earlier this year. Those close to the project said Apple concluded that Gemini offered the best blend of speed, reliability, and contextual understanding.

The revamped Siri, codenamed Linwood, is expected to launch next spring as part of iOS 26.4. The project, known internally as Glenwood, is being overseen by Mike Rockwell, the executive behind the Vision Pro headset, and software engineering chief Craig Federighi.

Under the terms being finalised, Google’s Gemini model will manage Siri’s “summariser” and “planner” functions, which help the assistant interpret user intent and coordinate complex actions. 

However, Apple’s own models will still handle several on-device tasks. To protect user data, Gemini will operate within Apple’s Private Cloud Compute servers rather than Google’s infrastructure.

Neither company has commented publicly on the partnership. Unlike the Safari search deal, Apple is expected to keep Google’s role behind the scenes, branding Siri’s improvements under its own ecosystem rather than sharing credit.

The collaboration is a rare moment of pragmatism from Apple, which has long avoided outsourcing key software capabilities. But as competitors like OpenAI, Microsoft, and Google grow quickly, Apple appears more willing to depend on outside systems, at least for now, to maintain competitiveness.

Despite leaning on Google, Apple has not abandoned its vision to build proprietary AI tools. The company’s in-house models team is reportedly developing a trillion-parameter cloud model, aiming to match Gemini’s quality by next year. Executives say they can phase out the Google technology in due course.

Globally, Apple is also preparing a version of the new Siri for the Chinese market, where Google services are banned. The Chinese variant is expected to run entirely on Apple’s own models with a compliance layer from Alibaba Group, tailored to meet local regulatory demands.

Shares of both companies briefly rose after reports of the talks surfaced, Apple gaining less than 1% to $271.70, and Alphabet rising as much as 3.2% to $286.42.