OpenAI is betting on audio. The company is developing a voice-activated AI device

OpenAI is focusing on audio format and preparing to launch a personal device with a voice interface.

This was reported by TechCrunch with reference to Information.

According to the publication, during November and December 2025, OpenAI brought together some engineering, product, and research teams to relaunch its audio models.

This move is part of a broader trend in Silicon Valley — the gradual abandonment of screens. Voice assistants have already become commonplace in more than a third of US households, TechCrunch writes. 

OpenAI is also working on its own device with an audio interface. It may be released in 2027. The company is considering creating a line of such devices — from screenless smart speakers to glasses.

Meta recently introduced a feature for Ray-Ban smart glasses that uses five microphones to help users hear their conversation partners in noisy environments.

Google is experimenting with “audio previews” of search results, and Tesla is integrating xAI’s Grok chatbot into its cars as a full-fledged voice assistant.

Apple is also working on creating its own smart glasses — Apple Glasses. According to the latest leaks, the new product could be unveiled as early as 2026. Unlike the Vision Pro headset, the new product will be more similar to the successful Ray-Ban Meta: without a built-in display, but with deep integration of artificial intelligence.

According to sources, OpenAI plans to release a new audio model in early 2026. It should sound more natural, support interruptions in conversation, and even speak simultaneously with the user — something that current models are unable to do. 

Read more: China’s ByteDance planed to spend $14 billion on NVIDIA AI chips in 2026