For a company that spent three years insisting it wasn’t in the hardware business, OpenAI keeps leaking details about a hardware business.
The latest comes from Bloomberg, which reports that OpenAI’s first consumer device — the one built with Jony Ive’s design studio, LoveFrom — will carry a price tag between $300 and $400. The shape, per earlier reporting, is a donut: a portable, screenless ring designed to move from kitchen counter to nightstand to home office, running on battery power and built from “high-quality metal” with a “premium look.” It reportedly has moving parts. Not buttons. Not a hinge. Parts that move on their own, apparently as a way of expressing something like personality.
Read those two details side by side, and the story stops being about a smart speaker and starts being about a bet on what a computer is supposed to feel like.
The Price Tag Is the Real Headline
Amazon’s Echo lineup runs from roughly $40 to $240. Sonos, at the premium end of the smart speaker world, tops out in a similar range. OpenAI’s device, if the reporting holds, will launch above nearly all of it — closer to a mid-range iPad than an Echo Dot.
That pricing only makes sense if OpenAI isn’t competing on the same axis as Amazon and Google. Those companies subsidized speaker hardware for years, treating the device as a loss-leading funnel into Prime orders and ad-supported search. Amazon has openly discussed years of losses on its device business. OpenAI, notably, has no equivalent flywheel to subsidize into. It has ChatGPT subscriptions and enterprise API revenue — neither of which needs a $350 donut to grow.
Which suggests the pricing isn’t a funnel strategy at all. It’s a positioning strategy. OpenAI wants the device read as a premium object, not a commodity appliance, and premium objects justify premium prices whether or not the underlying silicon costs more to make.
Moving Parts Are the Interesting Design Choice, Not the Screen’s Absence
Everyone covering this story has focused on the missing screen — understandable, since it’s a real departure from where Amazon, Google, and even Apple have been steering their smart-home hardware (increasingly toward displays, not away from them).
But the more telling detail is mechanical: a device engineered to physically move as part of how it communicates. Screenless devices already exist — plenty of smart speakers get by on light rings and chimes. Building actuated, moving components into a consumer electronics product is a much bigger manufacturing and durability commitment, and companies don’t take on that complexity by accident.
It reads like a deliberate attempt to give the object a presence — something closer to a pet or a companion than a speaker grille. That tracks with earlier reporting describing the device as OpenAI’s attempt at a “physical manifestation” of ChatGPT: not a tool you address, but something that seems to notice you back.
Whether that lands as charming or unsettling will depend entirely on execution most people haven’t seen yet. Anthropomorphized hardware has a mixed track record — think of how quickly novelty wears off on devices that try too hard to seem alive.
The Jony Ive Question Nobody’s Answered Yet
LoveFrom’s involvement does a lot of narrative work for OpenAI, and it’s worth being clear-eyed about why. Ive’s name signals “this will feel considered” in a way that reassures a market still skeptical that AI companies can build good physical objects — the same kind of goodwill OpenAI has struggled to manage on the softer side of its brand this year. But industrial design pedigree doesn’t resolve the harder question sitting underneath this device: does anyone actually want an object in their home that’s designed to seem like it has moods?
Smart speakers already answered a version of this question, and the market’s response was “somewhat, but not enough to pay much for it.” Alexa devices, by Amazon’s own admission, spent years running at a loss. Google Home and Nest went through consolidation and price cuts. The category exists, but it’s never proven especially lucrative — because voice assistants turned out to be genuinely useful for a narrow set of tasks (timers, weather, music) and much less useful as a standalone product people would pay a premium for.
OpenAI bets that a more capable model underneath — one that can hold a real conversation instead of parsing intents — changes that math entirely. That’s plausible. GPT-Live-class voice models are a meaningfully different experience than the scripted assistants Echo and Home shipped with in 2015. Whether “meaningfully better voice AI” converts into “worth $350 and a spot on your kitchen counter” is the actual experiment here, and it’s one no one has run yet at this price point.
The Lawsuit Sitting Underneath All of It
None of this is happening in a vacuum. Apple sued OpenAI in July, alleging the company used trade secrets — allegedly obtained through hires like Tang Tan, Evans Hankey, and Paul Meade, all former senior Apple hardware and design executives — to accelerate this exact device. OpenAI has denied wrongdoing and reportedly argues the product differs enough from Apple’s own hardware to sidestep the claims.
That fight matters for timing as much as optics. A 2027 release, per Bloomberg, was already an ambitious target for a company entering hardware manufacturing for the first time. An active trade-secrets suit involving the very designers building the product adds a variable that has nothing to do with battery life or actuator durability and everything to do with a courtroom calendar.
What This Actually Tests
Strip away the donut shape and the moving parts, and OpenAI is running a fairly specific experiment: whether people will pay hardware-premium prices for an AI presence rather than an AI utility. Smart speakers proved people wanted convenience at commodity prices. OpenAI is betting the next version of that category is companionship at Apple-adjacent prices — and that a genuinely better conversational model is the thing that finally makes the premium worth paying.
It’s a more interesting bet than “OpenAI makes a speaker.” It’s also a riskier one.
Related: OpenAI’s First AI Device Hits a $6.5 Billion Legal Battle With Apple
