Jony Ive’s first OpenAI gadget is reportedly a hockey puck-sized smart speaker

OpenAI is introducing a new hardware form factor that departs from the typical plug-in smart speaker. The device is described as a doughnut-shaped, battery-powered unit about the size of a hockey puck, emphasizing portability and one-hand operation. Its lack of a display means all user interaction will be voice-driven, supplemented by lights and moving parts that indicate activity. This design choice aligns with the goal of creating a “pocket personal assistant” that can be moved around the home easily.

Functionally, the speaker will run OpenAI’s more advanced conversational models, offering “humanlike interactivity” similar to the voice mode in ChatGPT mobile apps. In addition to microphones, it will house a camera system, lights, and other sensors, providing richer context for AI responses. The inclusion of moving parts that act autonomously to show when the device is responding adds a tactile feedback layer not common in current smart speakers. Engineers will need to account for these multimodal signals when designing interaction flows.

From a deployment perspective, the hardware is expected to cost over $300 at launch, establishing a higher upfront investment than many existing smart speakers. Developers will also need to integrate with OpenAI’s API endpoints, which may involve subscription fees and network latency considerations. Battery operation introduces constraints on continuous listening and processing, requiring efficient power management and possibly limiting always-on capabilities.

The speaker’s design imposes clear limits: without a screen, visual content delivery is impossible, and the reliance on battery power restricts runtime between charges. Its sensor suite, while broader than a typical speaker, still may not match the capabilities of full-featured smart displays or dedicated IoT hubs, potentially requiring additional hardware for comprehensive smart-home control. Compatibility with existing ecosystems (e.g., Alexa, Google Home) is not addressed, so integration may need custom bridges or adapters.

OpenAI frames this device as the first in a “family of devices,” suggesting future variants could expand functionality or form factor. For engineers, the launch signals a shift toward more physically distinct AI endpoints, prompting a reevaluation of how voice-first services are packaged and delivered. Early adoption will likely involve prototyping with the announced hardware specifications and testing the interaction model under the constraints of battery life and limited UI.