OpenAI reportedly picks MediaTek over Qualcomm for its maiden AI smartphone

OpenAI's first AI smartphone may enter mass production in 2027, with reports suggesting it could use a customised MediaTek chip and focus on agent-based, task-driven interactions

OpenAI has reportedly fast-tracked the development of its first artificial intelligence smartphone, which was earlier expected to enter mass production by 2028. TF International Securities analyst Ming-Chi Kuo said in a post on X that mass production could now begin as early as the first half of 2027.

 

Last month, Kuo had said that OpenAI was working with chipmakers MediaTek and Qualcomm, along with China-based Luxshare, for its AI-first phone. The latest update suggests that the company may have chosen MediaTek as its semiconductor partner, likely using a customised version of the Dimensity 9600 chip.

What the AI phone could offer

According to Kuo, the device may be powered by a customised MediaTek chip, possibly based on the upcoming Dimensity 9600. The chip could be built on TSMC’s next-generation N2P process and may be introduced in the second half of 2026.

 

The phone is expected to focus heavily on AI capabilities. One key area is the image signal processor, which may include an improved HDR pipeline to help the device better understand and analyse real-world scenes in real time.

 

The device could also feature a dual-NPU setup to handle different AI workloads more efficiently. It may use next-generation memory and storage, including LPDDR6 RAM and UFS 5.0, to support faster performance and smoother processing.

 

On the security side, the phone is expected to include features such as pKVM (protected virtualisation) and inline hashing to improve data protection and system integrity.

Shift towards AI-first interaction

One of the biggest changes could be in how users interact with the device. Instead of opening apps manually, users may be able to describe tasks, with the AI agent handling execution in the background.

 

The interface may focus on tasks rather than app icons. For instance, users could see updates on bookings, reminders or daily activities, along with progress indicators. Apps may still exist, but could run in the background without direct user input.

 

The device is also expected to combine cloud-based and on-device AI. This means it may process most tasks locally for faster and more private responses, while relying on cloud systems for more complex operations.

 

If development stays on track, Kuo estimates shipments could reach around 30 million units across 2027 and 2028, a significant target for a first-generation device.

 

Why OpenAI is entering smartphones

There is no official confirmation, but Kuo suggested that OpenAI may be aiming to control both hardware and software to deliver a more integrated AI experience.

 

Currently, tools such as ChatGPT operate within iOS and Android ecosystems, where system limitations can restrict functionality. A dedicated device could allow AI systems to perform tasks more directly.

 

Kuo also noted that smartphones provide continuous, real-time user data, which can help improve AI context and usefulness. With billions of devices in use globally, smartphones remain the largest distribution platform for AI services. 

Broader hardware ambitions

OpenAI is also reportedly exploring other AI-focused hardware products in collaboration with Jony Ive.

 

Reports suggest its first device could be a palm-sized, screen-less gadget that uses audio and video input to respond to user requests, possibly projecting information onto surfaces. The device is expected to remain always on and interact through a camera, microphone and speaker.

 

The company is also said to be exploring products such as smart glasses, a digital voice recorder and a wearable AI pin, indicating plans to build a broader ecosystem of AI-first devices.