OpenAI’s first consumer hardware product is no longer a rumor. According to Bloomberg’s Mark Gurman, the first device to come out of the OpenAI–Jony Ive collaboration is a hockey-puck-sized, ring-shaped smart speaker — no screen, priced around $300 to $400, expected to ship in 2027. In plainer terms: it looks like a metal donut that moves.
OpenAI internally refuses to call it a speaker. The official positioning is grander: the first computer built for AI — the physical embodiment of ChatGPT. That framing turns a product leak into something bigger: a bet on the entry point of human-computer interaction. When AI gets good enough, do we still need to stare at a screen all day?
Anatomy: A Donut That Moves
The product details are unusually specific for something this far from launch:
- Form factor: ring-shaped, roughly puck-sized, liftable with one hand and movable between rooms; battery-powered.
- No screen — but a speaker, microphone, lights, camera, and a suite of ambient sensors.
- A moving mechanical structure. Internal components physically shift during conversation while the lights change, signaling that the device is listening, responding, or executing — a deliberate effort to make it feel alive rather than sitting inert until summoned.
- Continuous perception. The camera and sensors mean it doesn’t just hear what you say; it can register what’s happening around you — the pill bottle on the table, the pot boiling on the stove.
- Personalization. It reads your email and other personal information, learns continuously, and aims to become the expert who knows you best — anticipating needs before you voice them.
The design is a direct answer to the failure mode of every smart speaker so far: Alexa and Google Assistant light up when called, answer, and go silent. What OpenAI wants is a persistent context — vision, voice, long-term memory, and environmental awareness fused into a presence that lives in your home continuously rather than conversing in disconnected turns.
The Roster: An Apple All-Star Team
The second layer of drama is the people behind it:
- Jony Ive. The designer of the iPhone, iPod, iPad, and Apple Watch — the man who defined consumer electronics’ product language for two decades. His studio LoveFrom handles the design.
- io Products. Ive’s hardware company, acquired by OpenAI in May 2025 for roughly $6.5 billion, with its ~200-person team absorbed outright — turning partnership into deep binding.
- Tang Tan. OpenAI’s hardware chief, formerly in charge of iPhone product design at Apple.
- Evans Hankey. Apple’s former head of industrial design, directly leading the speaker’s development.
- Paul Meade. Previously responsible for Apple Vision Pro and future smart-glasses development; OpenAI recruited him last month.
Apple’s response has been a lawsuit. The company alleges OpenAI obtained and used trade secrets covering product design, manufacturing processes, and supply chain — including asking Apple’s own suppliers to develop similar metal-finishing techniques — and counts more than 400 former Apple employees now at OpenAI. OpenAI denies the allegations and insists it is building “something completely different from Apple’s existing products.”
The more interesting split is strategic. Apple’s next-generation home device, by report, is a screen attached to a mechanical arm that can turn to face you. OpenAI is betting on removing the screen entirely. One company is pushing displays into more corners of the home; the other is testing whether, given a strong enough AI, the screen needs to exist at all. The symmetry is almost historical: Jony Ive helped design the touchscreen that swallowed consumer electronics twenty years ago, and now he is helping design a computing device that tries to swallow the touchscreen back.
AI Plus vs. AI-Native: Why the First Device Has No Screen
The smartphone’s core value for fifteen years has been built on screen plus apps: find information through the screen, invoke services through apps, complete every step through touch. The entire mobile internet — its business models, product design, advertising economics — is constructed on that interaction loop. Agents are dissolving it from underneath. When users only need to express intent, and the AI decides which software to call, which information to read, and which actions to take, the interface loses its job, and the screen loses its workload.
Here lies the distinction that explains the design:
- The AI+ logic adds a layer of intelligence to existing devices and existing interactions. Humans still open tools; tools still respond passively. Most current AI hardware — smart glasses, AI phone cases — walks this path.
- The AI-native logic refuses the patch and redefines how humans relate to computing. The model is the interaction; the device stops being an entrance and becomes part of the environment.
When a paradigm holds, hardware stops being a container for software and becomes the first-order constraint that shapes interaction itself. Every platform shift in computing history has moved the entrance: the PC era’s entrance was the desktop and the browser; mobile’s entrance was the app icon and the touchscreen; the AI era’s entrance is shifting from clicks to conversation, from operating systems to intent systems.
The cautionary tales are real, though. Humane Ai Pin proved that removing the screen does not, by itself, create the next computing platform — when model capability, response latency, agent execution, and product reliability can’t keep up, the missing screen amplifies every interaction failure. What separates OpenAI from those startups is that it already owns the models and the users; it is only now shopping for a physical body that fits them.
A Platform-Level Gamble in Three Layers
Stretched over time, OpenAI’s hardware ambition reads as three layers: enter the home as a speaker first, sidestepping the privacy minefield that wearables occupy (a continuously worn camera drags bystanders into its sensing range — the smart-glasses lesson already proved that technically feasible does not mean socially acceptable); then validate always-on interaction; and ultimately build a product with the power to replace the smartphone.
The industry consequences are structural. The smart-speaker category gets re-priced: Alexa and Google Home have been trapped in the voice-assistant box for years, and OpenAI is redefining the category as AI companion plus ambient computing — at $300 to $400, aimed straight at the premium end. Apple mounts its defense on two fronts — the lawsuit, and its own screen-toting home device — putting the two giants on a collision course over whether the next screen should exist, a question that will shape consumer hardware for a decade. And for AI hardware startups like Hark, Brett Adcock’s personal-intelligence venture that raised $700 million at a $6 billion valuation in May, OpenAI’s entry raises the bar for the entire field by an order of magnitude.
The risks are just as large: the privacy boundary of home sensing, the reliability and cost of moving mechanical parts, and the possibility that “presence” is a marketing gimmick rather than a feature. Three to four hundred dollars for a more expensive smart speaker — if agent capability can’t carry the scenarios, this becomes the next Humane.
Three Signals Worth Watching
- The maturity of real-time voice. The device’s core is natural conversation, so the ceiling of OpenAI’s real-time speech capability is the ceiling of the product. Positioned as ChatGPT’s physical form, model progress transmits directly into hardware experience.
- The privacy design for ambient context. Camera plus email access plus long-term memory is the combination that regulators, families, and the industry will scrutinize — its privacy architecture will become the reference standard for the entire AI-native hardware category.
- Release cadence and pricing. Between now and 2027, supply-chain leaks, price adjustments, and the progress of Apple’s lawsuit will keep rewriting the narrative.
Many of technology’s truly important products began as questions that sounded unnecessary. Why must the computer sit on a desk? Why must the phone have a cord? Why must the phone have a screen? This time, OpenAI is asking: when AI is strong enough, why must there be a screen at all? The questions everyone thought were pointless are usually the ones that move the industry.
