April 29, 2026, 7:15 am | Read time: 3 minutes
For some time, there have been indications that OpenAI is not only developing chatbots and AI models but also planning its own hardware. Initially, unusual devices were considered, including a pen-like input device as a direct interface to AI. Later, under the code name “Sweetpea,” a potential AI audio device came into focus, reminiscent of in-ear headphones, intended to integrate voice commands more into daily life. Now, the company is reportedly working on its own smartphone.
Planned Development with Partners
According to analyst Ming-Chi Kuo, OpenAI is collaborating with Qualcomm and MediaTek on the project. The focus is on a custom-developed smartphone chip specifically designed for AI requirements. Luxshare might also be involved. The Apple supplier is expected to handle system design, joint development, and manufacturing. Additionally, OpenAI is apparently leveraging internal expertise, such as the startup “io,” acquired in May 2025 for around $6.5 billion, founded by former Apple designer Jony Ive.
ChatGPT Creators Plan Their Own Alternative to X
Too Risky? OpenAI Suddenly Halts Its Erotic Chatbot
New Concept Instead of Classic Apps
The potential device would not be a traditional smartphone with a ChatGPT app. The focus is on a changed operation. Users would need to open fewer individual apps or navigate menus. Instead, an AI agent could take over tasks, linking information and managing processes. The smartphone would serve more as a central control unit.
For implementation, however, the technology would need to be more focused on AI functions than conventional smartphones. The device should recognize what is needed at the moment without heavily draining the battery. It is still unclear how OpenAI plans to achieve this. Possible approaches could include frequent access to the camera and microphone or analyzing screen content. However, this also raises privacy concerns, as more contextual data would be processed.
Also interesting: Too risky? OpenAI suddenly halts its erotic chatbot
Mix of Device and Cloud
Technically, a combination of local processing and cloud usage is planned. Simple tasks could be performed directly by the smartphone, such as providing short answers, summarizing texts, or recognizing connections. This would reduce response times and relieve external servers. For more complex tasks, however, OpenAI’s cloud would still be responsible, providing additional computing power.
It will take time before such a device could appear. According to Kuo, mass production is not expected to start before 2028. Key technical specifications could be established by the end of 2026 or early 2027. Until then, the project remains a potential glimpse into OpenAI’s plans in the hardware sector.