Why Apple Chose Google to Power its AI Revolution

Apple and Google's multi-year collaboration – costing the former US$1bn per year – will see Gemini models used for an AI-powered Siri.
Previously, Apple's Siri was considered to be under the category of 'narrow AI' – a type of AI which is designed for specific tasks.
However, Google Gemini 2.0 is a form of agentic AI and is multimodal, meaning it processes text, images and audio. It was made available to all users in February 2025. The model is more advanced than Siri because it has highly-advanced conversation skills, integration and personalisation skills.
Apple's aim for partnering with Google is to improve its user experience, harnessing the very same technology that has established Google as a true AI pioneer.
In a joint statement, Google and Apple said: "Apple Intelligence will continue to run on Apple devices and Private Cloud Compute." While Apple Intelligence will still be used, the focus is on developing Siri into a highly-intelligent agentic model, like Google Gemini 2.0.
Responsible AI
A hot topic of conversation surrounding Apple and Google's collaboration concerns how advanced AI models will be developed responsibly.
Ranjita Das, AI Product Manager at Stealth Startup, says: "Responsible AI is no longer about a policy you add at the end. It is an architectural decision.
"Responsibility should be assigned to system layers, not abstract principles. The layer closest to the user must retain control, override paths and accountability.
"With this move, Apple and Google once again show why they have repeatedly been the pioneers in shaping entire ecosystems."
AI amnesia
Most AI integrations currently lack "institutional memory", as Amar Ratnakar Naik, Vice President of Engineering at TELUS Digital, points out.
He describes the '10% barrier', where basic automations perform well, such as AI drafting, sending or rewriting emails. However, as users try to train AI over time and build on its knowledge, it begins to forget prior tasks and loses the ability to respond in the way required. As a result, success rates fall below 10%.
Currently, AI agents act as "contractors" rather than "employees", says Amar. In other words, they do the minimal work required for the amount of energy it uses. The AI lacks memory and doesn't learn over time, carrying out tasks but forgetting them the next day.
The partnership between Google and Apple aims to solve the institutional memory problem at grassroots-level. Through App Intents, one agent learns a new capability, which then becomes available system-wide.
The 'definitive corporate strategy'
Vikas Goel, Generative and Agentic AI Leader at PA Consulting, says the launch of DeepSeek-R1, a leading competitor to ChatGPT, demonstrated that the core infrastructure for AI is already in place.
He explains that the user experience layer has become significantly easier and cheaper to replicate.
According to Vikas, this is largely due to advances in AI integration, where a large "teacher" model can be used to transfer its learned patterns to a smaller "student" model, delivering near-equivalent performance at a fraction of the cost.
He suggests that, through their partnership, Apple is "outsourcing the foundational reasoning layer to Google", meaning Apple can use Google Gemini as a "teacher" to apply its model to Siri.
He describes this as the "definitive corporate strategy for 2026".
Tech's big players up the ante
Apple and Google's collaboration marks a clear shift for the former away from owning every layer of its technology, highlighting how its cautious in-house AI development has left it dependent on an external partner with greater scale and capability.
For Google, the deal embeds its AI deeply into Apple’s ecosystem, extending Gemini’s reach beyond Android and strengthening its position in the consumer AI race.
While AI has not yet been a decisive factor for iPhone buyers, analysts expect growing consumer demand, making this collaboration critical to Apple’s future relevance as AI-powered services become mainstream.




