OpenAI is preparing to take a delicate but strategic leap: going from being an AI software provider to competing in the smartphone marketAccording to several reports from analyst Ming-Chi Kuo, Sam Altman's company is already working on its own phone which, on paper, is aimed directly at the iPhone's territory and the hegemony of iOS and Android.
What's striking is not just that OpenAI wants to launch another mobile phone, but that the project is designed from the ground up to... Agentic artificial intelligence is the core of the device And not just a simple add-on. The proposal involves questioning the classic app model and suggesting a phone that operates around tasks, context, and agents that act on behalf of the user.
A smartphone designed for AI agents, not app icons
According to the information gathered by Kuo, OpenAI is defining a mobile phone where the main interaction will not be opening applicationsbut to ask for thingsThe user would formulate goals or needs in natural language, and the AI agents would break down those requests and execute actions using different background services.
In a typical scenario, instead of jumping between the calendar app, email, and a video calling tool, It would be enough to tell the phone to organize a meetingThe system would look for gaps in the schedule, propose a time, notify the contacts involved, reserve a room or restaurant, and generate a summary with the points to be discussed.
This approach requires the terminal to have a ongoing understanding of the user's contextLocation, activity, communications, recent history, and preferences. This is precisely the kind of information AI agents need to be truly useful and not just answer random questions like most assistants do today.
That's why Kuo emphasizes that the smartphone remains, to this day, the best-positioned device for the next generation of personal artificial intelligenceIt's always on the table, in your pocket or in your hand, with access to sensors, camera, microphones, payments and practically the entire digital life of the user.
Following one of the analyst's posts, Sam Altman hinted in X that it's a good time to “seriously rethink how operating systems and user interfaces are designed”He didn't mention any product by name, but the comment fits perfectly with the idea of a mobile phone where AI replaces the icon-based paradigm.
MediaTek, Qualcomm and Luxshare: the industrial alliance behind the project
To shape this device, OpenAI does not limit itself to buying standard components. Kuo claims that MediaTek and Qualcomm are working with the company on a specific processor for this smartphone.designed from the outset to support always-on AI agents.
It remains to be seen whether it's a completely new SoC or a deep adaptation of existing platforms, but the idea is clear: combine sufficient local computing power to understand the user's context with strong cloud support For the most demanding tasks. It's the same balance already seen in high-end mobile phones, but taken a step further and with AI at the heart of the design.
One of the key points will be energy consumption. The chip will need to be able to continuously process signals from the environment, language, and activity without draining the battery mid-day. Therefore, the processor design would prioritize memory management, efficiency in small form factors, and the ability to activate only the necessary blocks based on context.
In the area of assembly, the chosen partner would be Luxshare Precision. According to Kuo, Luxshare would become the exclusive manufacturer and co-designer of the phone, a high-profile contract for a company that has been gaining ground in Apple's supply chain for years with products like the iPhone and AirPods.
Mass production, according to the leaks, It wouldn't start before 2028The final hardware specifications, chip details, and the complete list of suppliers would be finalized between the end of 2026 and the first quarter of 2027. It's a long timeline, typical of a complex and large-scale project.
A strategic shift: from smart speaker to mobile phone that competes with the iPhone
What this smartphone proposes is a significant change of course from OpenAI's initial hardware plansThe company had initially paved the way for a home speaker with a front-facing camera, glasses with integrated AI, a connected lamp, and various accessories such as headphones.
These products are the result of the acquisition of io Products, the hardware startup founded by Jony Ive, Apple's long-time chief designerThe acquisition, valued at around $6.500 billion, led to a new division within OpenAI specifically focused on physical devices with artificial intelligence.
The original plan, publicly confirmed by executives like Chris Lehane, involved announcing that first smart speaker at the second half of 2026 and bring it to stores in early 2027. From there, the glasses, the lamp and the headphones would arrive, setting up a family of connected devices around ChatGPT.
Kuo's information doesn't contradict that timeline, but it does add a much more significant piece of information: a complete smartphone under the OpenAI brandThis phone wouldn't necessarily replace the rest of the products, but would expand the focus to the one device that almost no one is willing to leave at home.
Altman had previously suggested that the devices he was working on with Ive would be “different from smartphones”They compared the ideal experience to going from walking through Times Square, full of noise and stimuli, to sitting peacefully in a cabin by a lake. The mobile leak doesn't invalidate that vision, but it translates it into a format that most users understand and use daily.
An agentic model to challenge the iPhone ecosystem
Beyond the hardware, the core of the project lies in the software. OpenAI wants AI agents to replace much of the current role of applications.Instead of opening a transport app, a maps app, and a messaging app, the user would talk to the mobile phone and it would coordinate the necessary services in the background.
According to the sketches Kuo is working with, the device's main screen would be structured around a workflow instead of an icon gridThe user would see which tasks are in progress, which have been completed, and what remains pending: from booking a flight or paying for insurance to preparing a daily news summary.
The system would be organized into several basic sections, with names like Home, Actions, Memory, and Inbox. Each would group different types of information and It would allow AI to remember relevant context over time, something that is lost today when jumping between isolated applications.
This approach clashes head-on with the Apple's business model and, to a lesser extent, from Google. A large part of Cupertino's revenue comes from the app store and associated servicesIf a significant number of users were to stop interacting with specific apps and instead delegate most actions to agents, the role of the App Store would inevitably be reduced.
In Europe and other markets with stricter regulations, this type of phone may fit the current climate, in which Regulators are scrutinizing the dominant position of large platforms.An operating system designed around agents, if it transparently opens up space for developers, could also be sold as a less dependent alternative to the traditional closed garden.
Why OpenAI also wants to control the hardware
At the heart of it all is a question of dependency. Today, ChatGPT lives within iOS and Android with limited permissionsIt cannot freely access the microphone in the background, nor move around the system performing actions without going through security layers decided by Apple or Google.
For an assistant who intends to act in real time on the user's digital lifeThese restrictions are a significant obstacle. Hence, Kuo's reasoning suggests that OpenAI sees control of the operating system and hardware as an almost obligatory step if it wants to take its agentic model to its extreme.
Furthermore, the company has seen firsthand what happens when there is excessive reliance on third-party platforms. Apple has opted for its own models and for agreements with Google Gemini to strengthen SiriMeanwhile, Android is pushing for the integration of Google's AI into the system itself. In that scenario, ChatGPT risks becoming a secondary option.
Building their own mobile phone allows OpenAI to have a direct route to users without going through third-party rulesThat doesn't mean abandoning current apps, but rather creating a channel where the experience can be much deeper, even if sales volume doesn't approach that of the iPhone in its early years.
The movement also opens the door to new business formulasKuo doesn't rule out the possibility that the smartphone will be sold bundled with AI subscriptions, integrating ChatGPT's paid plans or premium services from day one. In that case, the device would be both a mobile phone and a gateway to an ecosystem of recurring services.
Impact on the market and medium-term uncertainties
The leaks have already had visible effects on financial markets. Qualcomm shares have surged in pre-market trading in New York. after its participation in the project became known, while Luxshare has seen its value rise sharply on Asian stock exchanges.
Apple, meanwhile, has suffered slight declines on the stock market in sessions following the publication of the reports, a sign that Investors interpret OpenAI's potential mobile phone as a medium-term threatespecially if it manages to capture some of the high-end market dominated by the iPhone.
Even so, nobody in the industry is under any illusions: Entering the smartphone market is extremely complicatedMicrosoft, Amazon, and Meta have already failed in previous attempts, even with resources and agreements with operators. Loyalty to existing ecosystems and the complexity of the supply chain work against new players.
OpenAI's differentiating factor would be that it doesn't intend to compete solely on technical specifications, but on change the usage paradigmIf the agents can solve real-world tasks better and faster than traditional apps, the phone could find a niche even if its initial sales aren't massive.
The big question is time. With mass production targeting 2028, The window of opportunity may narrow if Apple and Google accelerate their own integrated AI proposals. on iOS and Android. Between now and then, we'll likely see many advancements in contextual assistants on current systems.
In any case, the project sends a clear message: The next big battle will not only be over who has the best AI model, but over who controls the device where that AI lives.And OpenAI has decided that it doesn't want to continue playing only on other people's turf.
Based on the available information, everything points to OpenAI's future mobile project being a serious attempt at Redefine the smartphone as a screen-based assistant rather than a collection of applicationsIt's powered by a custom-designed chip from MediaTek and Qualcomm, assembled by Luxshare, and features an agent-centric usage model. It remains to be seen whether the market, accustomed to icons and app stores, is ready to make the leap, and, above all, whether OpenAI can translate its software advantage to the much more slippery terrain of consumer hardware.
