OpenAI Reportedly Building a Moving AI Speaker for the Home

According to a report in Bloomberg by Mark Gurman, the still-unannounced product is being developed as a humanlike AI companion capable of controlling smart-home appliances, playing media, answering questions and responding to messages. OpenAI reportedly sees it less as a conventional speaker and more as a new type of home computer built around artificial intelligence.

The device is said to include a camera and other sensors to understand its surroundings, along with a rechargeable battery so it can be carried between rooms. It may also feature moving mechanical elements, helping it feel less like a static gadget and more like a physical companion.

Its defining feature, however, is expected to be personality. The speaker would reportedly become more personalised and proactive over time, learning about its owner, anticipating needs and drawing on information such as emails to provide more relevant help.

That ambition also raises an obvious concern given a device using microphones, cameras, sensors and personal communications to build a detailed understanding of its owner will need to make a convincing case around privacy, security and user control.

Although it resembles a smart speaker, OpenAI appears to be reversing the usual formula. The Apple HomePod 2 combines music playback with Siri, HomeKit and Matter integration, while Sonos puts spatial music and Dolby Atmos at the centre of its Era 300.

Devialet’s Mania brought portability, adaptive processing and claimed 360-degree stereo sound to the category. More recently, StereoNET’s review of the WiiM Sound found a flexible, audio-led smart speaker offering hi-res streaming, voice control, multi-room playback and AI RoomFit optimisation.

In each case, the intelligence supports the speaker where as OpenAI’s reported device would instead use sound as one of several ways for an AI system to interact with its owner, with audio performance potentially taking a secondary role.

Its conversational abilities are expected to rely on GPT-Live, the voice technology OpenAI introduced this month. The full-duplex system can listen and respond continuously, following pauses, interruptions and changes in conversational pace rather than waiting for a clearly defined turn to end.

OpenAI’s move into hardware has been expected since it acquired io Products, the device startup co-founded by former Apple design chief Jony Ive, in a deal reportedly worth US$6.5 billion.

Ive’s independent design studio, LoveFrom, is also helping shape the wider product line. The project involves numerous former Apple designers and engineers who worked on products including the iPhone and Mac, with OpenAI reportedly hoping that experience will help make its first device feel more personal and approachable.

Some Apple connections have also created complications for OpenAI. Apple filed a lawsuit against OpenAI and two former employees in July 2026, alleging that confidential hardware information was improperly obtained and used to accelerate OpenAI’s device plans. OpenAI has denied having any interest in Apple’s trade secrets, and the allegations remain unproven.

OpenAI is reportedly aiming to announce the speaker before the end of 2026, followed by a commercial release in 2027, although the development process and legal dispute could affect that timing. Plenty remains unknown, including the driver arrangement, amplification, supported streaming services, smart-home standards, local processing, subscription requirements and price.

Siri, Alexa and Gemini have already made voice assistants familiar in the home. OpenAI’s first device appears to be pushing the idea further, testing whether people want ChatGPT to become a more proactive, personalised and increasingly observant presence in the room, rather than an assistant that waits to be called upon.

Join the Discussion