According to industry analyst Ming-Chi Kuo, OpenAI would accelerate the development of its first “AI agent phone”, with the company now aiming to mass produce the device as early as the first half of next year.
Late last month, Kuo revealed OpenAI’s work on a smartphone, contradicting earlier reports that the company had no plans to enter the mobile market. Kuo said MediaTek and Qualcomm are the chosen chip partners and Luxshare Precision Industry is the exclusive manufacturing partner, with mass production planned for 2028.
The reasons for Kuo’s revised production target for the first half of 2027 would now include OpenAI’s planned IPO (i.e. a compelling hardware product could strengthen its story with investors if it went public) and intensifying competition in AI agent phones. Kuo says MediaTek appears “best positioned to become the sole processor supplier”, with the device configured to use a customized version of the Dimensity 9600, which will apparently be built on TSMC’s N2P node during 2H26.
The “key spec” of the device would be its image signal processor, featuring an improved HDR pipeline that improves real-world detection – or what the AI “perceives” through the camera. The phone will also use dual AI processors to handle different tasks (e.g. vision and language simultaneously), fast memory and storage, and security features to isolate processes.
“If development remains on track, combined shipments between 2027 and 2028 could reach around 30 million units,” says Kuo, who says that full control of the operating system and hardware is the only way for the company to provide a complete AI agent service. Kuo expects AI agents to change the way people interact with a phone, shifting the focus from launching individual apps to performing tasks within a seamless pop-up interface.
It’s not entirely clear where this leads to Jony Ive’s prototype non-telephone AI device. Shortly after its acquisition of Ive’s startup io Products in May 2025, OpenAI embarked on a marketing campaign of sorts to promote Ive’s first upcoming product for the company, describing it as a “third primary device” after a MacBook Pro and an iPhone. According to Sam Altman, CEO of OpenAI, it would be “the coolest technology the world has ever seen.” And best of all, it wouldn’t have a screen – because phones have screens, and Ive and Altman want to keep people away from them.
The original goal was to release the device later this year, but in November that roadmap was pushed back to “less than two years.” The latest news is that Ive’s first OpenAI device was revealed to be a smart speaker with a camera, scheduled for release in early 2027.
Other OpenAI products are reportedly in development, including smart glasses, a smart lamp, and potentially headphones, but the roadmap is believed to be further out for those. If one of these devices ends up launching, OpenAI will become a direct hardware rival against several Apple product lines – Apple is also rumored to be working on smart glasses, AirPods with cameras and an AI pendant, as well as a smart home hub with enhanced Siri capabilities.