SAN FRANCISCO – The first hardware product OpenAI is developing for consumers is a speaker without a screen that physically moves on its own, is designed to function as a personalized AI companion, and can access the owner’s emails as part of how it learns about them. Bloomberg reported the details Monday, citing people familiar with OpenAI’s plans.
The device has not been officially announced and carries no stated release date or price. What Bloomberg’s report establishes is that OpenAI has made specific decisions about its physical form. The product contains “mechanical elements that can move on their own,” distinguishing it from the stationary smart speakers Amazon, Google, and Apple have built over the past decade. A speaker that repositions itself in space implies a kind of attention to presence that passive listening devices do not attempt.
OpenAI designed the device to have a distinct personality. The product is intended to “feel like a companion and become a physical manifestation of OpenAI’s ChatGPT,” according to people familiar with its development. It learns proactively about the people it lives with rather than waiting for prompts, constructing a personalized model that grows over time. Email access is part of that learning mechanism. The device would eventually hold a working knowledge of your inbox, your schedule, and your habits, not because you asked it to, but because it was built to accumulate that picture independently.
That integration sets a different threshold from what existing home AI products attempt. Alexa and Google Assistant respond to explicit requests. They answer questions, play music, control thermostats. None of them were designed to read your correspondence or build behavioral profiles without ongoing instruction. OpenAI’s described device would operate differently, forming a model of who you are across time and deploying that model in how it presents itself inside your home.
Former Apple engineers who worked on iPhone and Mac hardware assisted in developing the device, TechCrunch reported Monday, citing the Bloomberg findings. That detail carries significant legal weight given the current state of Apple’s federal trade secret lawsuit against OpenAI.
OpenAI addressed the legal dimension preemptively. The company said the device “veers significantly from anything Apple has on the market today” and called it “unlikely that it violates trade secrets” belonging to Apple. The affirmation is notable precisely because nobody asked. A company entirely comfortable with its legal position would typically describe a new product without volunteering an assessment of what it does not infringe.
Apple filed its trade secret complaint earlier this month, accusing Tang Tan, the company’s chief hardware officer and a 24-year Apple veteran who oversaw product design for the iPhone and Apple Watch, of orchestrating a scheme in which job candidates were coached to bring Apple hardware prototypes to interviews. The complaint alleges the misconduct was normalized through multiple levels of OpenAI’s senior leadership. OpenAI has denied wrongdoing. An attorney email mix-up that preceded the lawsuit’s filing has further complicated the narrative around how Apple pursued informal resolution before going to court.
The hardware program has proceeded through io, the design studio founded by Jony Ive, the former Apple chief design officer whose three decades at the company produced the iMac, the iPod, and the iPhone. OpenAI acquired io last year for $6.5 billion, making Ive’s team its principal hardware capability. Ive is not named in Apple’s lawsuit. The acquisition gave OpenAI access to design expertise that had defined the dominant consumer hardware category of the past generation.
The AI hardware market is still searching for its defining category. Meta’s Ray-Ban glasses reached consumers without a screen and found an audience. Amazon’s Astro home robot introduced autonomous physical movement into a living room device. Google’s Nest ecosystem remained passive and stationary. None of these products produced a breakout moment comparable to what smartphones achieved, and none was designed primarily to initiate interaction rather than respond to it. OpenAI’s described companion sits differently on that map, built to behave proactively rather than reactively, to move rather than sit, and to learn rather than look up.
What is not in Monday’s reporting is what the device physically looks like, how its mechanical movement presents itself to someone living with it, or when OpenAI expects to bring it to market. These are not minor omissions. The gap between a product described through anonymous sourcing and one available on a shelf is wide enough that reported details may evolve considerably before any release.
What is established is that OpenAI is building consumer hardware, that former Apple engineers are helping build it, that Apple is suing over what it says happened when those engineers left, and that the device would interact with users in ways no current home AI product attempts. Each of those facts is independently significant. Taken together, they describe a product arriving at a moment its own development history has made unusually complicated.
A screenless speaker that moves on its own and wants to know your email is either a genuinely new kind of AI product or an interface bet that will look misplaced in retrospect. OpenAI has apparently decided which one it is.
