What was announced
Apple used its annual developer conference on Monday, June 8, 2026, to unveil Siri AI, the most significant overhaul of its voice assistant in fifteen years, rebuilt from the ground up on a custom Google Gemini model. The keynote also marked Tim Cook's final appearance as CEO before he hands the role to hardware chief John Ternus on September 1, 2026.
Apple unveiled a rebuilt Siri powered by a custom 1.2-trillion-parameter Google Gemini model, alongside a developer preview of a new operating system called homeOS, and developer betas for six platforms: iOS 27, iPadOS 27, macOS 27, watchOS 27, tvOS 27, and visionOS 27.
Bloomberg reports that Apple will pay Google roughly $1 billion a year for the custom Gemini model that now powers the rebuilt Siri. That model is about eight times larger than the largest cloud model Apple had previously built in-house, and it uses a mixture-of-experts design — rather than activating all 1.2 trillion parameters per query, it routes each request to the relevant subset of specialised sub-networks, keeping latency competitive.
Why this is a genuinely unusual moment in Apple's history
To understand why this announcement matters, it helps to understand what Apple has always refused to do. For nearly two decades, Apple's defining strategic posture has been vertical integration — building its own chips, its own operating systems, its own software stack, top to bottom, rather than depending on a competitor for a critical layer of its product. This is the company that spent years and billions of dollars designing its own silicon specifically to avoid depending on Intel.
According to multiple reports, this decision is particularly revealing. For years, Apple favoured a fully internal approach. Now, the company appears to recognise that swiftly reaching the level of the top AI models sometimes necessitates strategic partnerships. Google, Meta, Microsoft, and a growing field of startups have been shipping conversational AI at scale for two years. Apple's Gemini deal effectively acknowledges that it could not build a competitive large language model in-house on the required timeline. What it offers instead is distribution, with more than two billion active devices, and a privacy architecture that none of its competitors can match.
The context that makes this decision even more pointed: Apple settled a $250 million consumer class action last month over marketing Siri AI features in 2024 that were not ready when the iPhone 16 launched. The personalised Siri capabilities originally advertised at WWDC 2024 were delayed indefinitely in March 2025. Monday's keynote was, in practical terms, the delivery of features Apple had been sued for failing to ship.
Apple tried to build this itself. It promised it publicly in 2024. It failed to deliver it. It got sued for the failure. And the way it resolved the lawsuit was not by finishing its own model — it was by licensing someone else's.
How the new Siri actually works
Siri now uses a three-tier routing system. Simple tasks stay on the device using Apple's own models. Moderately complex requests go to Apple's Private Cloud Compute servers. The heaviest reasoning tasks route to Google Cloud, running on Nvidia Blackwell B200 GPUs. At each step, Apple anonymises and tokenises queries so neither Apple staff nor Google can link requests to individual users.
The rebuilt assistant arrives as a standalone app on iPhone, iPad, and Mac, functioning as a conversational chatbot alongside its existing system-wide presence. It can draw on personal context to search across messages, emails, and photos, execute multi-step commands across apps, and answer questions about what is on screen.
Perhaps the most significant detail, for anyone tracking the broader AI industry, is this: iOS 27 Extensions will let users set a third-party AI model, such as Claude or ChatGPT, as their default assistant — a concession to both competitive pressure and the EU's Digital Markets Act requirements that previously forced Apple to pause its AI rollout in Europe. Apple is letting rival chatbots integrate with Siri in iOS 27, expanding on the OpenAI partnership that currently allows Siri to hand off requests to ChatGPT. Bloomberg reports Apple plans to allow other chatbots like Claude and Gemini to work with Siri, so users will be able to send questions to their favourite chatbot instead of Siri.
For the first time in the iPhone's history, the assistant baked into the operating system of two billion devices will not be a single company's product. It will be a menu.
What the keynote revealed about the state of the AI industry
Apple's chief software engineer used the same stage to draw a deliberate contrast with the rest of the industry. Craig Federighi argued that Apple Intelligence is more useful because it uses personal information and data, saying: "Some appear to be racing forward, seemingly pursuing AI for the sake of AI, without clear regard to the people, all of us, that it's ultimately meant to serve."
That line was almost certainly aimed at the same companies Apple just paid a billion dollars to access. It captures the tension running through the entire AI industry in 2026: a race to build the most capable possible system, set against a much harder, slower question about whether that system actually serves the person using it. Apple is betting that two billion devices and a credible privacy architecture matter more than building the smartest model in-house. Whether that bet pays off depends entirely on execution — something Apple has already failed at once in this same product cycle.
For developers, the more consequential story is that App Intents is now the mandatory way Siri talks to third-party apps, and SiriKit — the framework developers have used for a decade — received a formal deprecation notice, with Apple signalling a roughly two-to-three-year support window. Every app on the App Store now has a multi-year clock running on a mandatory AI integration migration. That is not a consumer story. It is an entire industry's engineering roadmap, set by a single keynote.
Why this matters for the children growing up with this technology
The most important lesson in this story is not about Siri, or Gemini, or even Apple. It is about what the announcement reveals about how AI capability actually gets built and distributed in the world your students will inherit.
For years, the assumption in technology has been that the largest, most resourced companies can build anything they need in-house. Apple's decision says otherwise. Building frontier-level AI capability — the kind that can power a conversational assistant trusted by two billion people — requires resources, talent, and research infrastructure so specialised that even one of the most valuable companies in history concluded it was faster, cheaper, and more reliable to license it from someone else. That is a genuinely new dynamic in the history of consumer technology, and it tells your students something important about the world of work they are entering: expertise in AI itself — not just the ability to use a chatbot, but the deep technical capacity to build, train, and deploy frontier models — is now one of the scarcest and most valuable skills on the planet. Scarce enough that Apple pays roughly $1 billion a year rather than compete for it.
It also tells them something about choice. iOS 27 will let a thirteen-year-old choose whether Siri, Claude, or ChatGPT answers their questions by default. That is an extraordinary amount of power to place in the hands of a child who has never been taught how to evaluate which AI system to trust, how their data moves between Apple's servers and Google's, or what a three-tier routing architecture actually means for their privacy. The technical literacy required to make that choice well is not a niche skill anymore. It is becoming a basic requirement of using a smartphone.
This is precisely the gap AI Ready School's programmes are built to close. Cypher teaches students to question what an AI tells them, not just receive it — a habit of mind that matters just as much when a thirteen-year-old is choosing between three different AI assistants on their own phone as it does in a classroom. NEO gives children hands-on experience with how AI systems are actually built and trained, so that a billion-dollar licensing deal between two of the largest companies on Earth is not an abstract business story to them, but something they have a genuine technical intuition for. The AI Workshop for Teachers and the Agentic AI for Kids programme exist because the assistant on every student's phone just changed — fundamentally, structurally — and the adults responsible for guiding those students need to understand the change at least as well as the students will.
The detail worth remembering
Tim Cook ended the WWDC keynote with a short, personal farewell message: "Over the years, you have helped people connect, create, learn, and experience the world in extraordinary new ways. And with the incredible capabilities we introduced today, and so many more still to come, I truly believe the best is still ahead."
Cook delivered that line in the same keynote where Apple admitted it needed a competitor's AI to deliver on a promise it made two years earlier. The best may well still be ahead. But it is worth noticing that even Apple — the company that built its identity on doing everything itself — concluded that getting there required a different kind of partnership than the one its reputation was built on. The schools paying attention to that shift, and preparing their students for a world defined by it, are the ones that will matter most in the years it describes.
