OpenAI is shifting its AI strategy to audio and is working on a personal AI device with a voice interface, TechCrunch reported, citing The Information.
According to the report, OpenAI began rebuilding its audio models in November-December 2025 by merging several engineering, product, and research teams. This move is seen as part of the “screenless technologies” trend that is gaining momentum in Silicon Valley. Currently, more than a third of households in the United States use voice assistants.
OpenAI is also developing its own voice AI device. According to sources, this device can be introduced in 2027. The company is considering releasing devices in various formats, from screenless smart speakers to smart glasses.
Note that Meta, Google, Tesla, and Apple are also actively working on similar audio-based technologies. OpenAI is expected to introduce a new audio model with more natural speech and real-time dialogue capabilities in early 2026.
