Apple is shipping its rebuilt assistant, Siri AI, as a beta product wrapped in daily usage caps, a possible waitlist, and gaps in language and regional coverage, the company confirmed Wednesday. The assistant reaches the public on September 14 inside iOS 27, iPadOS 27, macOS 27 (code-named Golden Gate), watchOS 27, and visionOS 27, roughly two years after Apple first previewed a personalized Siri as part of Apple Intelligence at its 2024 developer conference.

The stakes are straightforward: this is Apple’s answer to the personal-context, cross-app assistants that Google and OpenAI have already shipped, and it is arriving with more asterisks than either competitor attached to a comparable launch. A feature Apple promised would define its AI strategy is going out the door still labeled unfinished.

Apple’s release notes single out four features for daily server-side limits: Image Playground, Siri AI itself, the Apple Foundation 3 cloud models that Shortcuts calls on, and intelligent photo editing. The caps shift by feature, by how complex a given request is, by how much load Apple’s servers are under, and by policies the company has not detailed. Apple says an expanded, paid tier is coming, but it has disclosed neither a price nor a launch date, an unusual gap for a company that typically ties feature announcements to a specific ship date.

Regional and language coverage is uneven in ways that will matter more to some users than others. Siri AI launches in English only, with French, Japanese, Korean, Portuguese, and Spanish scheduled for October 2026. Mainland China is excluded entirely at launch. The European Union split is the more striking case: Apple is withholding Siri AI from iPhone, iPad, and Apple Watch there while permitting it on Mac and Apple Vision Pro, a platform-by-platform carve-out that suggests regulatory friction specific to mobile rather than a blanket EU block. Apple has not stated which rule is driving the iPhone exclusion. Users under 13 are locked out regardless of device.

Code AppleInsider found in August described a waitlist mechanism that would give users an estimated wait time tied to server capacity and update it as demand shifted. Apple has not said whether that system activates at launch, which leaves open whether “beta” here means limited features or a metered rollout similar to how OpenAI and Anthropic throttled early access to compute-heavy products.

On paper, the assistant matches what Apple promised in 2024: natural conversation, open-ended question answering, and retrieval of personal context across email, notes, messages, and photos without the user naming where the information lives. It can act inside Messages, Music, and Reminders based on what is on screen, and a dedicated Siri app now stores conversation history across devices. Visual search through the Camera app, customizable voices, and drafting or revising text round out the feature list, though several of those depend on the same server capacity that is already rationed.

Apple’s decision to gate a flagship AI feature behind usage caps while still withholding pricing puts it in the same monetization bind as the API providers it is chasing: use limits without a clear self-serve upgrade path tend to push power users toward whichever competitor charges upfront but does not throttle. Enterprise IT teams evaluating Apple Intelligence for managed fleets should treat the October multilingual rollout, not the September 14 date, as the real readiness marker.

Reporting by Andrew Orr for AppleInsider, published September 9, 2026.