According to MacRumors, reporting on a Bloomberg story by Mark Gurman, Apple is preparing to release two new versions of Siri despite losing at least four AI researchers and a high-ranking Siri executive, Stuart Bowers, who joined Google’s DeepMind. The first update, a more personalized Siri powered by Google Gemini, is slated for iOS 26.4, which should enter beta in February and launch publicly in March or April 2025. This version will understand personal context and have on-screen awareness. The second, more ambitious update planned for iOS 27 will turn Siri into a full-fledged chatbot capable of sustained conversation, built directly into Apple devices. Gurman says this Siri chatbot aims to be “competitive with Gemini 3” and far more capable than the iOS 26.4 version.
The roadmap is clear, but the exodus is louder
Here’s the thing: the technical roadmap Gurman lays out actually makes a ton of sense. A phased rollout—first a smarter, context-aware Siri, then a full conversational agent—is the logical, Apple way to do it. It avoids dumping a half-baked ChatGPT clone on users all at once. But the personnel news completely overshadows the product announcement. Losing a senior exec like Stuart Bowers and four researchers isn’t just “some employees” leaving; it’s a brain drain to a direct competitor, Google. That tells a story no slick launch timeline can erase. It suggests internal friction, or perhaps that Apple’s infamous secrecy and vertical integration is frustrating top AI talent who want to publish and push boundaries faster.
The Gemini paradox
So Apple is using Google’s Gemini to power the first phase of its Siri revival. Let that sink in. The company that prides itself on controlling the whole stack is outsourcing the core AI for its flagship virtual assistant to its arch-rival. That’s a huge, humbling admission. It basically says, “We can’t do this part ourselves yet, and we need to ship *something*.” It’s a pragmatic, maybe even smart, short-term move. But long-term? It’s awkward. It cements Google’s AI as the industry benchmark that even Apple has to license. And it raises a big question: will the “super Siri” in iOS 27 finally be powered entirely by Apple’s own models, or will that partnership with Google deepen?
What this means for your phone
Forget the corporate drama. What does this mean if you’re holding an iPhone? Basically, Siri is finally going to stop being a punchline. By late 2025 or 2026, it should be able to actually help you in a real, multi-turn conversation. It’ll understand what’s on your screen and what you’re trying to do in an app. That’s the dream we were sold over a decade ago. The catch is timing. iOS 26.4’s update is still months away, and the chatbot version is likely a year and a half out. In the hyper-fast world of AI, that’s an eternity. By the time Apple’s in-house chatbot Siri arrives, what will ChatGPT and Gemini be doing? That’s the execution risk Apple is taking with this measured, two-step plan.
