OpenAI Wants to Build a ‘Third Device’ to Break Your Screen Habit
OpenAI is attempting to move beyond the chatbot. The company is accelerating development on a screenless, audio-first hardware device designed to function as a proactive digital assistant, aiming for a consumer launch in 2026.
According to recent reporting from The Information, the San Francisco-based startup has reorganized its engineering teams to prioritize a new generation of audio models. The goal is to create a “third core device” that sits alongside the smartphone and laptop, relying on voice rather than visual interfaces.
Instead of a tool that requires opening an app, the hardware is being pitched internally as a persistent presence.
“Think of it less as a tool you open, and more as a presence that’s just there. A companion that notices patterns in your life and tries to help.”
The ‘Always-On’ Concept
Prototypes currently under discussion range from iPod-like handhelds to smart glasses and pen-shaped wearables. The common denominator is the removal of the friction associated with unlocking a phone. Executives envision a device that utilizes microphones, cameras, and speakers to process natural speech and context in real-time.
This aligns with a broader industry shift toward “ambient computing,” a concept pursued by Apple and Meta, where technology recedes into the background of the home or wardrobe.
Solving the Latency Problem
To make a voice-only device viable, OpenAI is overhauling its audio infrastructure. Current voice assistants often suffer from lag and robotic cadence. Over the last two months, OpenAI has consolidated research and product teams under a mandate to build models that mimic human conversational speed and fluidity.
The upcoming model, slated for early 2026, aims to solve specific technical hurdles: speaking with emotional inflection, responding instantly, and handling interruptions gracefully—a capability known in the industry as “barge-in” support.
“If it feels like a call center script, people will turn it off. The goal is to make it feel like a really sharp friend on the other end of the line.”
A Shift from Reactive to Proactive
The philosophical shift at OpenAI is significant. Rather than waiting for a prompt, the device is designed to be proactive. Internal presentations have highlighted use cases where the AI suggests leaving early to beat traffic or offers real-time translations during a conversation in a foreign language.
This proactive model attempts to bypass the “memory friction” that plagues current AI tools—the requirement that users must remember to ask the AI for help.
The Jony Ive and Foxconn Connection
OpenAI is bolstering its lack of hardware DNA through high-profile partnerships. The company acquired io, the design firm co-founded by former Apple design chief Jony Ive, in a deal valued at nearly $6.5 billion. Ive, who has frequently criticized the addictive nature of modern smartphones, is leading a team of over 50 designers to shape the device.
On the manufacturing front, OpenAI is reportedly in talks with Foxconn to produce the device in Vietnam, signaling an intent to scale production quickly upon launch.
The Privacy Implications
The utility of a proactive assistant relies heavily on surveillance. For the device to offer relevant suggestions without being prompted, it must effectively listen—and potentially watch—continuously. This raises immediate privacy concerns that go beyond the scope of a text-based chatbot.
Natalia Dominguez, a policy analyst specializing in biometric surveillance, notes the cultural hurdle.
“You’re talking about something that sits on your body or in your home and can capture not just you, but everyone around you. We’ve already struggled with smart speakers in living rooms. This is that, but mobile and more intimate.”
OpenAI has not yet clarified how it will handle voice data retention, local versus cloud processing, or bystander privacy. These factors are likely to attract scrutiny from regulators in Europe and states like Illinois, which have strict biometric privacy laws.
Market Skepticism
While the project has high-profile backing, the market for screenless AI hardware is littered with failures, including the Humane Ai Pin and the Rabbit R1. Skeptics argue that while consumers want smarter assistants, they may not be willing to abandon the utility of a screen or tolerate the privacy trade-offs of an always-listening microphone.
Court filings related to the io acquisition suggest the product design is not yet finalized. With no prototype publicly revealed and a launch date years away, OpenAI faces a long road to prove it can control the device in your pocket as effectively as it controls the models in the cloud.
