OpenAI is reportedly building its first device, and it is not a phone. According to Bloomberg, it is a screenless smart speaker designed with Jony Ive, small enough to carry from room to room, somewhere between $200 and $300, with a camera, environmental sensors, moving parts meant to make it feel alive, and ChatGPT for a brain. Target: 2027. The Wire flagged it this week, and it earned the longer conversation, because if even half the report is true, this is the most interesting thing to happen to the home speaker since the first Echo showed up and started setting kitchen timers.
I want to walk through what is actually being claimed, what it could mean for getting real intelligence into our houses, and the privacy question that comes with it.
What Do the Rumors Actually Say?
Here is what has been reported. Bloomberg says that OpenAI’s first hardware is a portable, screenless speaker developed under Jony Ive, the man who designed the iPhone, through his LoveFrom studio. It runs an advanced version of ChatGPT’s voice mode. It answers questions, plays media, responds to messages, and controls smart home gear. It has a camera and sensors, so it can see the room, not just hear it. It runs on a battery, so it follows you around the house instead of living on one counter. And the detail everyone keeps repeating, it has motorized elements that move on their own, because it is designed to feel less like a gadget and more like a companion.
The reporting goes further. The device is supposed to get more personal and more proactive the longer you own it, learning your routines and preferences, and drawing on your personal information, including email, to build that understanding. It is apparently one of roughly five devices OpenAI has in the pipeline, with a long-term ambition of building something that replaces the smartphone.
Now the caveats. This is a report, not an announcement. The timeline has already slipped once, from 2026 to 2027. OpenAI has never shipped a piece of consumer hardware in its life. And there is active litigation with Apple hovering over the whole project. We wrote a whole section of the vision essay about companies wheeling magical home companions onto stages and never shipping them, and the patron saint of that routine was a rolling ball from a company that has been building hardware for fifty years.
One more piece belongs in the recap, and it is the reason for this article’s title. At OpenAI’s developer conference last fall, Ive told the audience he does not think “we have an easy relationship with our technology at the moment,” and that the goal for these devices is to “make us happy, and fulfilled, and more peaceful, and less anxious, and less disconnected.” Sit with who is saying that. The man who designed the iPhone built the thing none of us can put down. Now he says he wants to build the thing that lets us put it down. Whether that is redemption or just very good marketing, it is the right ambition for technology that lives in your house. The gear that earns a permanent place in a home is the gear you stop noticing.
Why Is This Different From the Speaker Already on Your Counter?
Because the speakers we own today are remote controls with a voice.
I say that with affection. Alexa and Google Assistant are good at the unglamorous work, lights off, timer set, garage closed. But they execute commands. They do not understand your house, and they certainly do not understand your life. Every ounce of “smart” in today’s smart home is smart because somebody, usually me, on a Saturday, wired the logic together by hand.
The pitch here is a different animal entirely. A device with a frontier AI model for a brain does not need me to program the morning routine. It is supposed to notice the routine. It sees that the kitchen lights come on at six, that the thermostat gets nudged before the school run, that the house goes quiet at ten, and it starts running the pattern before anyone asks. Then February brings a new work schedule, and instead of me re-wiring a dozen automations, the house just… adapts. Learns the new rhythm. Adjusts as life adjusts.
That is not an incremental feature. That is the actual dream, technology that recedes because it handles itself. What has been missing between a schedule that fires on time and a house that actually knows you is intelligence, and the honest reason it has been missing is that no assistant in our homes has had any. A speaker that puts a genuinely capable AI in the middle of the house, with eyes, ears, and a memory for patterns, is the first plausible bridge across that gap. The industry has been inching this direction all year, the whole CES conversation shifted from AI features to AI coordination, but this would be the first device built for that idea from scratch rather than bolted onto a 2014 speaker design.
And because it reportedly controls existing smart home gear rather than replacing it, your current switches, plugs, and sensors become the hands. The speaker becomes the brain. If it speaks Matter properly, everything you have already installed is suddenly a limb of something much smarter. That is a big if, and I will get to it.
What About the Privacy Question?
There are two versions of this question.
The reflexive version says an always-listening speaker with a camera that reads your email is a surveillance device, full stop. The ingredients are all there. But here is the uncomfortable math. You already carry a microphone, a camera, GPS, and your entire correspondence in your pocket, everywhere, all day. There is probably already a listening speaker in your kitchen. The camera-and-microphone ship sailed years ago, and pretending this one device is the line we will not cross is theater.
The adult question is not whether a device can listen. It is who processes what it hears, where that processing happens, what gets kept, and what the business model wants with it. That is the lens that matters, and we just watched two companies answer it in opposite ways. Apple built its new camera intelligence to run on your own hardware wherever possible, with a cloud designed not to remember. Blink shipped AI features wearing a monthly fee, processed in the big cloud, with the guardrails coming from state legislatures rather than the company.
Where would OpenAI land? Unknown, and worth being clear-eyed about. Frontier models mostly live in data centers, which means this device would likely be the most cloud-dependent thing in your house, and a device that improves by reading your email is, by definition, a device that reads your email. None of that is disqualifying. All of it belongs on the label. The standard I will hold it to is the same one I hold everything to, a house that pays attention to you, not one that files reports on you. OpenAI gets to prove which one it built.
Would It Actually Replace Your Hub?
Here is my honest handicapping, three years out from a rumored ship date.
The bear case writes itself. Hardware is brutally hard, ask the fifty-year-old companies that have missed this exact target. The 2027 date has already slipped once. A lawsuit is circling. And the smart home graveyard is full of devices that promised to be the one brain your house needed and became ecosystem number five instead. If this thing arrives speaking only its own language, demanding its own plugs and bulbs, it is dead on arrival no matter how charming the motorized parts are.
The bull case is simpler and, I think, stronger. Somebody is eventually going to put real intelligence in the middle of the home. Every incumbent is retrofitting old assistants to fake it. OpenAI would be starting from the intelligence and building the house around it, with the best industrial designer of his generation shaping the thing. Even if this exact device slips or stumbles, it drags the entire category toward the home that learns, and Amazon, Google, and Apple will not sit still while it does.
My advice does not change either way. Do not wait for it, and do not pre-order the future. Keep building the boring, reliable base, the setup that already works, on open standards a future brain can plug into. Every dumb-but-dependable switch you install today is a limb that some genuinely intelligent thing will eventually be glad to have. For the full Bloomberg report, the details are here.
Until then, I remain what I have always been, an optimist with a scar tissue collection. The house that learns is coming. Whether it arrives in 2027 wearing a Jony Ive shell is the part I would not bet the mortgage on.
FAQ: The OpenAI Home Speaker
According to a Bloomberg report, OpenAI’s first device is a portable, screenless smart speaker designed with Jony Ive. It runs ChatGPT’s voice mode, includes a camera and sensors, controls smart home devices, and is designed to learn your routines over time. OpenAI has not officially announced it.
The reported target is 2027. The timeline has already slipped once, from an original 2026 goal, and the device could be delayed further.
Reports place the price between $200 and $300, in the same range as premium smart speakers from Amazon, Google, and Apple.
The report says it will control smart home gear, but OpenAI has not confirmed which standards it will support. Whether it works with Matter and existing ecosystems is the single biggest open question for real smart homes.


