AI smartphones bid farewell to the feature race; Honor begins reconstructing the operating system.

The competition among AI smartphones is moving beyond the stage of simply adding functions.

On July 18, during the 2026 World Artificial Intelligence Conference, Honor released the Agentic OS technology framework. According to Honor's vision, the next-generation mobile operating system will shift from being application-centered to being centered on the user's "intent" and "task."

In the past two years, smartphone manufacturers have mainly competed based on features such as AI erasing, call summaries, and document summarization. Although large models have entered smartphones, users' operating methods have not fundamentally changed: they still need to open the relevant app, then gradually complete searching, selecting, and confirming.

Honor hopes to further transform this interaction logic, which has been used for many years.

Huang Fei, Honor's chief AI scientist, stated that the essence of Agentic OS is not "adding an AI assistant to the system," but restructuring a new operating system centered on "intent" and "task."

Users no longer need to first determine which app to open, but only need to express their end goal. The system is then responsible for understanding the intent, breaking down the task, planning the path, and invoking different apps and services to complete the execution. For example, when a user wants to order a cake, take a taxi to a restaurant, and book a KTV room, the phone needs to recognize that the user is preparing a birthday party, then break this goal into several steps, and continuously complete operations across different apps.

Apps thus gradually recede into the background, and user intent becomes the new entry point. The smartphone also shifts from being a container that waits for users to operate each app, to a platform capable of organizing models, apps, and hardware to complete tasks.

However, moving from "understanding a sentence" to actually "completing a task" is not just a change in interaction method—it also requires the operating system to have a sufficient supply of models behind it.

Besides continuing to develop its own large models for terminal scenarios, Honor also announced this time a partnership with Alibaba’s Qianwen, jointly developing deeply optimized large model solutions for smartphone scenarios.

This cooperation also reflects the realistic choices of smartphone manufacturers in the era of large models. It is difficult for a terminal manufacturer to independently train and maintain all types of models and to cover all tasks users might propose.

In the future, competition in AI smartphones may not depend on whether vendors have fully self-developed large models, but on whether they can organize self-developed models, external models, app services, and hardware capabilities into a stable system.

If Agentic OS solves how to orchestrate models and apps in the digital world, then Honor's Robot Phone further extends this capability to the physical world.

At the WAIC venue, Honor demonstrated some of Agentic OS's capabilities through Robot Phone. Users can initiate cross-app tasks in natural language, and the system automatically completes task breakdown, path planning, and execution.

When there are deviations during task execution, the agent can trace back previous steps and adjust the operation path; for risky operations such as payment, the system will pause automatic execution, hand control back to the user, and protect sensitive information through screen blackouts and other methods.

These capabilities primarily verify whether the phone’s agent can complete tasks continuously like a human and can autonomously correct mistakes after errors.

On this basis, Robot Phone adds a four-degree-of-freedom gimbal system. It can recognize and follow the user's movement, and also move to the rhythm of music.

From cross-app operations to gimbal control, Robot Phone serves more as a technology testbed: verifying whether, once the operating system can unify the scheduling of models, apps, sensors, and actuators, the phone can further transform from a digital tool into an intelligent terminal with perception and action capabilities.

Honor’s judgment about the future of smartphones is also based on this logic.

Honor believes that in the next five years, smartphones will not be completely replaced by new AI hardware like glasses or earphones, but their positioning will gradually shift from being the unique hub of personal devices to the core node connecting different terminals and agents.

As single-point features like AI erasing and summarization become standard configuration for smartphones, AI smartphone differentiation starts to move down to the operating system level.

What product form AI smartphones will ultimately take still has no definitive answer. However, moving from function stacking to system restructuring has already become an important path for smartphone manufacturers seeking the next-generation interaction entry point.

Risk warning and disclaimer clauseThe market carries risks. Investment requires caution. This article does not constitute individual investment advice, nor does it take into account the particular investment objectives, financial status, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article suit their specific circumstances. Invest accordingly, at your own risk.