Tim Cook took the stage at Apple Park one last time Monday morning, and the centerpiece of his farewell keynote was a confession dressed up as a product launch: Apple cannot build a competitive AI assistant on its own.
The new Siri, rebuilt from the ground up and powered by a custom 1.2-trillion-parameter Google Gemini model, is the most dramatic concession to a rival that Apple has made in decades.
The Billion-Dollar Handshake
The deal, first reported by Tom’s Guide, gives Apple a licensed Gemini model running as the backbone of Siri’s cloud intelligence at a reported cost of roughly $1 billion per year. That is not a licensing fee. That is a surrender premium. Apple spent years and billions trying to catch up to OpenAI and Google on large language models, and the Gemini partnership is the clearest signal yet that Cupertino decided buying was cheaper than building.
The new Siri looks nothing like the old one. It now operates as a full chatbot with a dedicated app, a system-wide “Search or Ask” gesture, and integration into the Dynamic Island on iPhone 16 and newer. Users can type or talk. Multi-step commands work. Siri can compose emails, pull personal context, read what is on screen, and chain actions across apps.
If that sounds familiar, it should. This is the interface paradigm that ChatGPT and Claude established two years ago. Apple is not leading here. Apple is catching up, and the fact that it needed Google’s model to do it tells you everything about where the AI race actually stands.
iOS 27 and the Platform Updates
Beyond Siri, the keynote rolled out iOS 27, iPadOS 27, macOS 27, watchOS 27, and visionOS 27. TechRadar’s live coverage noted deeper Apple Intelligence integration across the board: natural language photo editing, generative fill in the Camera app, pro-level customizable camera controls, and an improved system-wide grammar checker.
The Camera and Photos updates are where Apple’s on-device AI strategy still shines. Running inference locally, without a cloud round-trip, preserves the privacy story that Apple has spent a decade building. The tension is obvious: the most capable new feature (Siri) runs on Google’s cloud, while the privacy-first features run on-device with Apple’s own silicon. That split will define the next chapter of Apple’s platform story.
The Succession Angle
Cook announced in April that he will hand the CEO title to hardware chief John Ternus on September 1 and shift to executive chairman. This keynote was a deliberate bookend. Cook has led WWDC keynotes since 2012, and Monday’s event carried the weight of a farewell tour, complete with a standing ovation from the developer audience.
Ternus inherits a company that is simultaneously the most valuable on Earth and two years behind on the technology that investors believe will define the next decade. The Gemini deal buys time, but it also creates dependency. If Google raises the licensing price, if regulators in the EU or DOJ target the arrangement, or if Gemini falls behind a future OpenAI model, Apple is locked into a partner whose interests do not always align with its own.
That is the strategic risk Cook is handing to his successor, wrapped in a billion-dollar annual invoice.
What This Means for the AI Landscape
The Apple-Google Gemini partnership reshapes the competitive map. Microsoft has OpenAI. Apple now has Google. Meta is running Llama open-source. Amazon has Anthropic. The major consumer platforms have each picked an AI horse, and the era of any single company owning the full stack from silicon to model to interface is over.
For consumers, the immediate upside is real: Siri should finally work the way people have wanted it to since 2011. For the industry, the long-term question is whether Apple’s walled garden can survive when the intelligence inside it belongs to someone else.
Cook’s last keynote may have been his most consequential. Not because of what Apple built on its own, but because of what it admitted it could not.