Apple and Google are building AI together now. That’s the big news out of WWDC this week, and honestly, it’s stranger than it sounds at first.
We got our first real look at Apple’s AI strategy during the Worldwide Developers Conference, and it turns out the rumors were true: five new Apple Intelligence foundation models exist, built with Google’s infrastructure. But here’s the thing. You won’t be chatting with these models directly. They’re not the car. They’re the road.
What Exactly Are These Models?
Let’s unpack what Apple actually announced. The company released five new Apple Foundation Models, and they’re being positioned as the bedrock of the AI experience rather than the experience itself. Think of them as the underlying infrastructure that powers everything Siri does.
The new models include an on-device version and cloud versions, with the cloud variants apparently running on Google Cloud using Nvidia GPUs. Apple’s stance on privacy still applies, which is notable given that these cloud operations technically route through Google’s infrastructure. The company is sticking hard to its “we don’t store your data” messaging, even as it partners with a competitor on the actual AI grunt work.
Developers will be able to tap into these models through Apple’s Core AI framework, which opens up interesting possibilities for third-party integrations.
Siri is getting a major overhaul. The new assistant will be multimodal, meaning it can understand both speech and images, and it will have deeper access to your personal context across apps. Apple is promising better dictation, improved language understanding, and a system that can coordinate across multiple models. That last part is significant. Apple is essentially building an orchestrator that can pick the right AI tool for the job, rather than relying on a single monolithic model.
The On-Device Play
One thing Apple isn’t backing down on is on-device processing. The company is betting big on smaller models that run locally on your iPhone, iPad, or Mac rather than hitting the cloud for everything. It’s a privacy-first approach, sure, but it’s also a practical one. Latency matters, and nobody wants their voice assistant freezing up because it waiting for a server response.
This is where Apple’s strategy starts to look more considered than some of its competitors. While Google and Microsoft have been in a race to push bigger, cloud-heavy models, Apple is quietly building infrastructure that keeps more processing on your device. It’s less flashy, but it might actually matter more for everyday use.
The tradeoff is that these on-device models are necessarily smaller and less capable than what you can run on a data center. So Apple is essentially creating a two-tier system: lightweight tasks handled locally, heavier lifting sent to the cloud when needed.
Is This Revolutionary or Just… Iteration?
Let’s be honest about what we saw at WWDC. The upgraded Siri understands you better. Great. It can access more of your information to provide context. Useful. But revolutionary? That’s a harder sell.
Francisco Jeronimo, vice president of client devices at IDC, put it well in an email. He said Apple wants its AI to be “trustworthy and invisible to the user.” Instead of AI standing out as a feature, Apple is trying to weave it into the operating system so seamlessly that you barely notice it’s there.
That could be genuinely significant. If Apple can make AI feel natural, private, and actually useful for mainstream users, it might reshape what people expect from their devices. But we’re not there yet, and the gap between Apple’s vision and the current reality is still pretty wide.
The company has been relatively slow to the AI game compared to Google and OpenAI. Apple’s models haven’t gone viral. There hasn’t been a ChatGPT moment. Instead, we’re getting a gradual rollout of incremental improvements wrapped in heavy branding. That might work for Apple, but it’s hard not to feel like we’re watching a company that’s more reacting to the market than leading it.
What does seem clear is that Apple’s approach to AI is fundamentally different from its rivals. While everyone else is racing to build flashy chatbots and viral features, Apple is playing a longer game focused on integration, privacy, and infrastructure. Whether that pays off remains to be seen, but it’s certainly not boring.
