Apple is heading into WWDC 2026 with one of its biggest challenges in years. According to analyst Ming-Chi Kuo, simply catching up with competitors in AI is no longer enough. The new Siri needs to deliver more than just a "finally on par" experience—it must offer genuinely useful AI applications, agentic scenarios, and a hybrid approach blending on-device and cloud processing that surpasses Google Gemini.
Kuo argues that the real takeaway from WWDC won't be Apple's short-term stock reaction, but whether the company can create a more convenient and valuable user experience than Google itself—even while relying on Gemini for some underlying technology. This is especially critical after years of Siri updates falling short. Apple promised personalization, on-screen context awareness, and deeper integration with user data back in 2024.
Rumors suggest the new Siri in iOS 27 will have access to personal information across different apps, enabling complex multi-step commands. This capability is expected to come from App Integration, a system that extends the assistant's reach beyond Apple's own apps to third-party services. Combined with on-screen understanding and support for multiple large language models, Siri could evolve from a voice assistant into a personal AI agent.
The big question is how Apple will reconcile its privacy-first philosophy with the power of external models. According to Bloomberg, the future Siri might use an advanced version of Gemini, with some complex queries processed via Google Cloud on servers equipped with NVIDIA B200. Meanwhile, Apple is reportedly training smaller on-device models through distillation, bringing some Gemini capabilities directly to the device. If this combination proves faster, more private, and more useful than Google's own offerings, WWDC 2026 could mark Apple's return to the AI race—not as a follower, but as a potential leader.