Since Apple first released Apple Intelligence at WWDC in 2024, the launch date of the new AI version of Siri has been continuously delayed. According to Bloomberg reporter Mark Gurman published on February 11Latest reports, many important features originally scheduled to be launched with iOS 26.4 in March may now be split to iOS 26.5 (expected in May) or even iOS 27 (September) before they can meet users.
Apple’s new Siri development reportedly encounters bottlenecks! Some features may be delayed to iOS 26.5 or even iOS 27
According to Bloomberg, citing people familiar with the matter, Apple’s engineering team discovered new software issues during internal testing in recent weeks. The new version of Siri in testing sometimes fails to correctly handle user query requests, and the response speed is too slow, failing to meet the online standards. Apple has asked engineers to use iOS 26.5 instead as a target version for further internal testing.
A wait that starts in 2024
Looking back at this period of history, the fate of the new version of Siri can be said to be twists and turns:
- June 2024: Apple released Apple Intelligence at WWDC, showing off a smarter Siri (it was revealed that it only had a PPT demonstration effect)
- March 2025: Apple officially announced that the new Siri feature “will take longer to develop”, delayed by a full year
- January 2026: Apple and Google reached a cooperation agreement,Confirmed that the new Siri will use Google Gemini as the AI backend engine
- February 2026: Internal testing encountered obstacles again, and functions were forced to be postponed in batches.
Before selecting Google Gemini, Apple had also evaluated its own model as well as Anthropic and OpenAI’s solutions before finally deciding to cooperate with Google.
What exactly does the new Siri do?
According to MacRumors Break the news, currently in the internal beta version of iOS 26.5, all the features previously promised by Apple are included, but not all of them work stably. Key features include:
Personalized data access: Siri will be able to access users’ personal data such as text messages and emails. For example, you can ask Siri: “Help me find that podcast that a friend sent me before.” It will search it from the message history and play it directly. This feature is considered “most likely to be delayed”.
screen awareness: Siri can understand what’s currently on your screen and provide assistance accordingly.
Cross-App operations: Siri will be able to execute more complex commands between different apps and achieve deeper system integration.
Voice control in-app operations: Control functions within the app through voice commands, but this feature is currently lagging behind.
In addition, Apple is also testing two features that have not yet been officially released:
AI image generation: Integrate Image Playground to allow Siri to directly generate images.
Web search summary: Similar to Perplexity, Siri can search web pages and automatically organize summaries.
Both features are already in beta for iOS 26.4 and may still be rolled out with 26.4.
MacRumors also revealed an interesting detail: During testing, the new version of Siri would sometimes “skip” the Gemini engine and fallback directly to ChatGPT to answer questions. Even though the new Siri powered by Gemini is fully capable of handling the request, this means that Apple still has technical debt that needs to be resolved in the routing logic of the AI model.
What happens next?
Bloomberg pointed out that the current situation is still “fluid” and Apple’s plans may be adjusted again at any time. Executives don’t want to delay Siri functionality any further than spring 2026.
The most likely schedule at the moment:
- iOS 26.4 (March): May still include image generation and web search capabilities
- iOS 26.5 (May): Core functions such as personalized data access and screen sensing
- iOS 27 (September): Completely integrated with Gemini’s chatbot mode, allowing Siri to compete head-on with conversational AIs such as ChatGPT, Claude, and Gemini.
It has been nearly two years since the new Siri was unveiled at WWDC in 2024. The Siri in users’ hands is still the same as “often not understanding what you are saying.” Apple’s choice to cooperate with Google Gemini instead of conducting self-research reflects the anxiety in the AI race; now even integration is having difficulties, which highlights how high the technical threshold is for “deeply embedding LLM into the operating system.”
However, instead of rushing to launch a semi-finished product on the shelves, Apple chose to postpone polishing, which is the right direction. There is really not much time left for Siri to catch up. Other AI leaders such as ChatGPT, Gemini, and Claude have no intention of waiting for Apple. As for the recent rumors that the new Siri will be launched later in iOS 26.4, they may still have to wait.
Source: KOCPC Chinese