Apple Unveils Rebuilt Siri AI—Powered by Google, But Limited Global Rollout Sparks Questions

At WWDC 2026, Apple’s Stacey Ford, vice president of OS Program Management, remarked that “we’ve all had that moment where you search for something you know is there, but it just won’t show up.” While she was referring to Spotlight, the line also echoed Apple’s broader struggle with artificial intelligence.

That “missing” piece was finally revealed on Monday at Apple Park: the long-anticipated, fully rebuilt Siri AI. After years of falling short, Siri has been redesigned to support natural multi-turn conversations, access user data such as emails, messages, and photos, respond to real-time web queries, and execute actions across different apps. Apple is also launching it as a standalone app while keeping deep system integration, with iPhones displaying Siri activity through the Dynamic Island during ongoing requests. However, beyond the on-stage presentation, the more important story lies in the details—particularly what powers Siri AI and who can actually use it.

Google inside Apple’s AI stack

One of Apple’s most significant revelations was made almost in passing: the company confirmed it has partnered with Google and its Gemini model family to help develop the next generation of Apple Foundation Models, which underpin Apple Intelligence and now Siri AI. After years of insisting its own in-house AI would eventually catch up, Apple’s progress appears to have depended on external support.

Apple executives anticipated concerns about privacy. Senior vice president Craig Federighi emphasized that “privacy in AI is non-negotiable,” stating that user data is only used to fulfill requests and that independent experts can verify these claims. While the privacy framework may be solid, the strategic implications are harder to ignore. Apple is now relying on its biggest competitor in search and AI to power its assistant, even as Google expands Gemini across Android, Workspace, and its hardware ecosystem.

This effectively signals that Apple could not match frontier AI model development on its own timeline. If even a company with Apple’s resources, silicon advantage, and global scale chose to license rather than fully build its own models, it raises broader questions about how realistic “sovereign AI” ambitions are for other governments and companies.

Limited rollout raises global questions

The rollout plan for Siri AI also highlights major limitations. The initial beta, expected later this year, will only support English. China is excluded entirely due to unresolved regulatory issues, while users in the European Union will not receive the assistant on iPhone or iPad at launch. Apple says it is working on availability, but for now, EU access will be limited to macOS 27 and visionOS 27.

From a global perspective, the gaps are significant. China—one of Apple’s most competitive markets—is left out, while local AI assistants continue to operate freely. Meanwhile, users in major non-English-speaking regions across Asia, including Mandarin, Japanese, Korean, Hindi, and Bahasa speakers, will not have access to the new Siri experience for an unspecified period, remaining on the older version.

Apple has not provided a timeline for additional language support, marking a notable shift for a company known for simultaneous global product launches.

A quieter spotlight at WWDC

The structure of the keynote itself reflected Apple’s cautious positioning. As noted by TechCrunch, the presentation focused first on fixing existing issues before introducing new features, with Siri AI presented as just one part of a broader software update rather than the central highlight.

It also marked a leadership transition moment, with Tim Cook’s final WWDC as CEO ahead of John Ternus taking over Apple’s hardware leadership role on September 1. Cook closed by saying, “I truly believe the best is still ahead at Apple.”

Siri AI is now real and significantly more capable than before, showing Apple’s strength in integrating software across its ecosystem. But with core intelligence partially powered by external models and a rollout that excludes large parts of the global user base, the company’s “catch-up” in AI appears to have only begun.