Apple Opens the Door — iOS 27 Will Let You Choose Your Own AI

Apple is not trying to build the best AI model. It is trying to build the best AI platform. And with iOS 27, that strategy is about to become visible to every iPhone user on the planet.

The "Extensions" Strategy

Apple is reportedly preparing a major AI platform shift that would allow users to select third-party AI providers such as Google and Anthropic to power Apple Intelligence features across iOS 27, iPadOS 27 and macOS 27. The capability, internally referred to as "Extensions," would allow AI providers to integrate through App Store applications, giving users greater flexibility over which models handle text generation, editing and image tasks.
This is a fundamental repositioning. Until now, Apple Intelligence meant Apple's models — with a Google Gemini option for certain Siri queries. iOS 27 would open the entire Apple Intelligence layer to competition, letting users choose their preferred AI the way they choose their preferred browser or email app.

Why This Is Smarter Than It Looks

The move signals Apple's attempt to accelerate its AI competitiveness after lagging behind Microsoft and Google in consumer AI deployment. But framing it as a catch-up move misses the deeper strategy.
Apple controls the hardware. Apple controls the operating system. Apple controls the App Store through which every AI provider must distribute their Extensions. In a world where Google, Anthropic, OpenAI and Meta all compete to be the AI model inside your iPhone, Apple becomes the platform they all pay to access — and the company that sets the rules for how they behave on its devices.
Some leaders inside Apple think that large language models will be commoditised in the coming years, making internal model development difficult to justify. If they are right, the company that owns the distribution layer — not the model layer — wins. That company is Apple.

Meta's Parallel Problem

While Apple opens its platform, Meta is discovering the limits of its closed approach. Meta's next flagship model, internally called "Avocado," was pushed from March to May 2026 after internal testing showed the model underperforming against Google, OpenAI and Anthropic — particularly on coding and reasoning benchmarks.
Meta's AI division has seen significant restructuring, including leadership changes and the departure of several researchers who recently joined Meta Superintelligence Labs. Last month, chief AI scientist Yann LeCun announced he is leaving to start his own company.

Meta is spending $145 billion on AI infrastructure in 2026. Its flagship model is delayed. Its top AI scientist has left. And Apple — spending a fraction of that — is turning the entire AI model market into a feature of its platform.

What This Means for GAFAM

The iOS 27 Extensions strategy reshapes the competitive dynamics of the entire GAFAM AI race. Google and Anthropic now have an incentive to make their models work better on Apple devices — on Apple's terms. Microsoft's OpenAI relationship gives it a potential Extensions partner. Meta, without a consumer AI model competitive enough to attract iPhone users, faces the risk of being locked out of the world's most valuable device ecosystem.
Apple's AI push is also expected to expand into new product categories including AI-powered smart glasses and an expanded range of HomePod devices with more advanced home hub functionality. If those devices run on the same Extensions framework, Apple will have built an AI ecosystem that rivals compete inside — rather than against.

The European Perspective

The EU Digital Markets Act already requires Apple to allow third-party browser and app store competition on iOS. The Extensions framework raises a new regulatory question: if Apple controls which AI providers can integrate with Apple Intelligence — and on what terms — does that constitute a gatekeeping behaviour subject to DMA enforcement? Brussels will be watching iOS 27 very carefully. gafam.ai will be watching too.