- Apple and Google sign a multi-year agreement to make Gemini models the basis of Apple Foundation Models and the new Siri.
- The alliance accelerates Apple Intelligence's roadmap after delays, leadership changes, and problems deploying a truly advanced Siri.
- Apple promises to maintain on-device and on its private cloud execution, emphasizing privacy, while regulatory doubts grow in Europe and the US.
- The deal strengthens Google's power in AI, provokes a reaction from Elon Musk, and reignites the debate about technological consolidation.
The move many analysts considered a foregone conclusion is now official: Apple will leverage Google's Gemini processors to power the next generation of Siri and its Apple Foundation Models. What until recently sounded like an unlikely rumor has become one of the most significant technological alliances of the decade, with a direct impact on the future of the iPhone and the Apple ecosystem.
The Cupertino company had been under pressure for months due to its delays in artificial intelligence, while competitors like Google, OpenAI, and Microsoft boasted about their far more capable assistants. With this agreement, Apple effectively admits it needs a partner in generative AI to avoid falling behind , while insisting it will maintain control over the user experience, brand, and, above all, data privacy.
A historic agreement: Gemini as the basis for Apple Foundation Models
Apple and Google have announced a multi-year collaboration agreement under which the next generation of Apple Foundation Models will be built on Google's Gemini processors and some of its cloud infrastructure. In their joint statements, both companies describe the partnership as spanning several years and supporting upcoming Apple Intelligence features, including a more personalized version of Siri to be released later this year.
According to the two companies, the agreement comes after a “comprehensive evaluation” of various AI technologies, in which Apple reportedly concluded that the Gemini models provide “the most capable foundation” for its ambitions. In practice, this means that Google’s large language and multimodal models will become the backbone upon which Apple will build its advanced capabilities for language understanding, personal context, screen vision, and chained actions between apps.
From a technical standpoint, the integration goes beyond a simple external plugin: Gemini becomes the foundational layer upon which Apple will train and adapt its own models , with the goal of seamlessly integrating them into iOS, iPadOS, and macOS. For the end user, Siri will remain "Siri," and Apple Intelligence will continue to appear under the Apple brand; there will be no Gemini logos or visible references to Google in the interface.
The financial terms of the agreement have not been officially disclosed, but reports from outlets such as Bloomberg suggest an annual payment of around $1.000 billion from Apple for access to large Gemini models (configurations of around 1,2 trillion parameters have been mentioned, although without public confirmation). In any case, the agreement builds upon the long-standing business relationship between the two companies, which already includes the use of Google as the default search engine on Apple devices.
From WWDC to delays: why Apple has had to rely on Google
To understand the significance of this shift, it's helpful to recall the context. At WWDC 2024, Apple unveiled Apple Intelligence as its major response to generative AI , placing Siri at the heart of a strategy that promised an assistant capable of understanding the user's personal context, "seeing" what's on the screen, and chaining complex actions between applications. On paper, the leap from traditional Siri was enormous.
Throughout the second half of 2024, Apple maintained an optimistic tone. In December, the company reiterated in a press release that Siri's most advanced features would arrive "in the coming months ," while it began rolling out other components of Apple Intelligence, such as Image Playground, Genmoji, and visual analysis and assistive writing tools. The message was clear: the bulk of Siri's revolution was just around the corner.
However, as 2025 progressed, the tone shifted. In March, in a statement to the Daring Fireball blog, Apple acknowledged that some of these capabilities would require more development time than initially anticipated and began talking about a “more personalized” Siri that would be released “sometime next year.” The company itself admitted that it had a “version 1” ready, which was halted because it was deemed insufficient for internal standards and user expectations.
WWDC 2025 brought the problem to the forefront . Twelve months after promising a significant leap forward, the new Siri failed to deliver the level of sophistication Apple had hinted at. Craig Federighi, head of software, and Greg Joswiak, head of marketing, were forced to offer public explanations, emphasizing that they preferred to delay the launch rather than release an assistant that would fall short compared to what other market players already offered.
Internal changes, leadership crisis and a change of course in AI
The Siri debacle wasn't just a purely technical issue: it ended up having a major impact on Apple's organizational structure . In March 2025, agencies like Bloomberg and Reuters reported that the assistant was being removed from the area headed by John Giannandrea , the veteran recruit from Google who had led Apple's AI strategy since 2018.
From then on, direct responsibility for Siri passed to Mike Rockwell, known for his role in the development of Apple Vision Pro , who began reporting more directly to Craig Federighi. The internal message was clear: the software division needed to regain control of a product considered key to the platform's future.
Months later, on December 1, 2025, the company officially confirmed that Giannandrea would be leaving his position . After a transitional period as an advisor, he is scheduled to retire in the spring of 2026. At the same time, Apple appointed Amar Subramanya as the new vice president of AI, also under Federighi's leadership, tasked with redirecting the strategy in a context of increasing pressure.
In this scenario of broken promises, mounting delays, and internal restructuring , leaks about a possible alliance with Google began to gain traction. Throughout much of 2025, industry sources—including analyst Mark Gurman—indicated that Apple was exploring several AI providers, including OpenAI, Anthropic, and Perplexity, but that talks with Google were progressing more rapidly.
The officialization of the agreement in early 2026 confirms that Apple has opted for a pragmatic change of course: if it could not arrive on time with its own top-level models, the solution was to rely on one of the market leaders to accelerate the roadmap for Siri and the rest of Apple Intelligence.
What changes for Siri and Apple Intelligence with the arrival of Gemini
The most visible impact of the agreement will likely be the transformation of Siri into a much more capable assistant . Apple has placed an expanded version of Siri on its immediate roadmap, which should begin rolling out this year, tied to iOS 26.4 updates . The first developer beta could appear in a matter of weeks, with a gradual rollout throughout 2026.
Based on this, the company plans to launch new Apple Intelligence features that will arrive with later versions of the system, such as iOS 27. The idea is for the assistant to be able to interpret emails, messages, appointments, and files, combine information from different apps, and execute chained actions without the user having to manually switch between them.
In the demonstrations Apple has been showing, they've mentioned queries like, " What time does my flight leave and how long does it take to get to the airport from here? " or " Reschedule dinner with Marta if my meeting on Thursday changes ," using data from Mail, Calendar, Messages, and Maps. The ambition is for the new Siri to move beyond simply launching commands and become an intelligent orchestrator of everything that happens on the device.
Gemini fits into this vision by providing large-scale models with advanced natural language understanding, reasoning, and text and image generation capabilities. While Apple doesn't specify which exact Gemini variants it will use, it's assumed that it will combine lighter models for local tasks with larger models running in the cloud , all under the control of its own infrastructure.
All of this is in addition to the Apple Intelligence components already presented, such as the image and emoji generators (Image Playground and Genmoji), the visual analysis tools (Visual Intelligence), and the writing assistants. Furthermore, the option to access ChatGPT from Siri remains , although the focus of Apple's AI strategy now clearly shifts to Gemini.
Privacy and cloud: how Apple wants to maintain control
One of the most sensitive points of the agreement is the management of personal data and privacy , an area in which Apple has built much of its public image in recent years. With the entry of a third party as powerful as Google, doubts have quickly arisen, especially in markets with strict regulations like the European Union.
Apple and Google insist that, despite relying on Gemini's models and Google's cloud, Apple Intelligence will continue to run on Apple devices and its Private Cloud Compute infrastructure . In other words, processing will be done on the iPhone, iPad, or Mac whenever possible, and the cloud will only be used when the complexity of the request requires it.
In these cases, Apple assures that the data travels encrypted to servers under its control, specifically designed as a "secure zone" where not even the company itself can access the content of the requests. Google provides the modeling technology and some of the computing power, but Apple emphasizes that it maintains control over the architecture, including security safeguards and key management.
This approach also explains why only certain devices will be compatible with the most advanced features . Apple has stated on several occasions that running AI models locally requires a minimum level of processing power and memory, which in practice excludes a large portion of older iPhones. Currently, the cutoff is devices like the iPhone 15 Pro and later, which will have clear implications for European users when these features are activated.
However, Apple's commitment to vertical control is nuanced: for the first time, it is structurally relying on external models for such a central component as its assistant . The company is attempting to compensate with its own privacy layer and product design, but the balance between technological dependence and data protection will be subject to scrutiny, especially from European authorities.
Market reactions, Musk, and the regulatory spotlight in Europe
In financial markets, the announcement has been received as a positive sign for both Apple and Google . Shares of both companies rose slightly after the agreement was made public, and in Alphabet's case, the boost helped its market capitalization reach $4 trillion in intraday trading, joining the club already comprised of Nvidia, Microsoft, and Apple itself.
Analysts like Dan Ives of Wedbush Securities have described the alliance as “a major validation moment for Google” and “a springboard for Apple” in its attempt to catch up in artificial intelligence. The interpretation on Wall Street is that the deal reduces the risk of Apple falling behind its competitors and, at the same time, solidifies Gemini as one of the industry's benchmark models.
On the other side, one of the most critical voices has been that of Elon Musk . The founder of xAI and head of Grok reacted to the announcement via X (formerly Twitter), responding to Google's official post with a message stating that "this seems like an unreasonable concentration of power for Google, given that they also own Android and Chrome ." His comments align with the fear that a single entity could effectively control AI on the two major mobile platforms.
Musk's stance is not neutral: xAI competes directly with Gemini and OpenAI's models , and is also engaged in a legal dispute with Apple and OpenAI in the United States, which it accuses of coordinating to ensure their dominance in the AI market. In this context, Google's entry as Apple's preferred partner reinforces the feeling of isolation for alternative projects like Grok.
Beyond the disputes between giants, the agreement crosses paths with an increasingly demanding regulatory landscape, especially in the EU. The long-standing collaboration between Apple and Google over the default search engine was already challenged in court , with the US Department of Justice arguing that the pact reinforced Google's dominance in online search. Although the judge allowed the agreement to continue, the precedent weighs heavily now that both companies are joining forces again in such a sensitive area as AI.
For European authorities responsible for competition and compliance with regulations such as the Digital Markets Act, the combination of dominant operating systems, benchmark AI models, and control of cloud infrastructure could become a new area of scrutiny. It is possible that Brussels will request further explanations or specific conditions to ensure that other AI providers and mobile platforms can continue to compete on a level playing field.
Impact on users in Spain and Europe
On a practical level, iPhone and other Apple device users in Spain and the rest of Europe are wondering when they will actually see the effects of the agreement . For now, everything points to a gradual rollout: first, an improved Siri focused on understanding language and personal context, and then, more ambitious features tied to later versions of the operating system.
The company will also have to coordinate these launches with the European regulatory framework , including obligations arising from the Digital Services Act and future specific regulations on AI. Issues such as the processing of personal data, the possibility of choosing alternative models, and transparency in the use of external models will be key when introducing the new Siri in markets like Spain.
It's expected that, at least in the initial phase, not all European countries will receive exactly the same features at the same time , whether due to technical reasons, regulatory requirements, or strategic decisions by Apple. We've seen this before—for example, with health or financial services features—where the company adapts the rollout pace to the legal conditions of each region.
At the same time, the agreement could influence the adoption of new devices in Europe . If the most powerful capabilities of Apple Intelligence and the new Siri are restricted to recent models like the iPhone 15 Pro and later, many users might consider upgrading their devices sooner than expected to access all the new features. In a mature market, this upgrade incentive is not a minor detail.
It remains to be seen how Gemini will coexist with other models that Apple may integrate in the future. The company itself has hinted that the agreement with Google is not necessarily exclusive , which opens the door for different AI providers to coexist over time, depending on the type of task, the level of privacy required, or the user's preferences.
The alliance between Apple and Google marks a turning point in how we understand the relationship between major tech companies: two long-time rivals in mobile are joining forces in artificial intelligence to avoid falling behind other equally powerful players . For users in Spain and Europe, the results will be measured in their daily use, when Siri has to prove whether, with Gemini under the hood, it finally lives up to the expectations built up over the years.
