OpenAI’s long-anticipated move into hardware is finally taking shape, and it isn’t the iPhone rival many expected. According to Bloomberg’s Mark Gurman, citing people familiar with the project, OpenAI’s first consumer device will be a portable, screen-free smart speaker that the company describes internally as a new kind of home computer for the AI era. The pitch is deliberately not another Alexa: OpenAI wants it to function as a humanlike AI companion that lives in the home, learns its owner over time, and acts as a physical manifestation of ChatGPT.
What the Device Actually Does
On the surface it behaves like a smart speaker, controlling smart-home appliances, playing media, answering questions, and responding to messages through the full range of ChatGPT’s capabilities. Underneath, OpenAI is describing something closer to an always-aware ambient computer. It carries a rechargeable battery so it can be carried from the kitchen to the laundry room to the bedroom over the course of a day, or left plugged into one room. Its voice abilities run on GPT-Live, the more advanced ChatGPT Voice Mode OpenAI rolled out this month, which can listen and speak at the same time, interject with acknowledgment cues, and adapt mid-conversation to sound more human. The Information reports pricing is expected to land between $200 and $300, cheaper than Apple’s HomePod but above a basic Echo Dot, with a launch targeted for early 2027 and manufacturing planned in Vietnam or the US.
The Camera Is the Real Story
The feature that separates this from every smart speaker on the market is what it sees and how continuously it listens. The device includes a camera and additional sensors that let it identify objects on a nearby table, recognize the people in the room, and follow conversations happening in its vicinity. It’s engineered for persistent listening without a wake word, meaning it tracks ambient context continuously rather than waking only on “Alexa” or “Hey Google.” The camera also powers a facial-recognition system similar to Apple’s Face ID, which can authenticate purchases and unlock personalized profiles for different household members without a PIN. In an internal presentation, OpenAI reportedly told employees the speaker would observe users and proactively suggest actions to help them hit their goals, for example, nudging someone toward an earlier bedtime ahead of a morning meeting. Taken together, persistent listening, visual scene awareness, and biometric recognition form a capability stack no current smart speaker offers.
Built to Feel Alive
OpenAI believes the product’s defining feature will be its personality. The speaker incorporates mechanical elements that can move on their own, intended to create the sense that it’s alive rather than an object waiting for commands, though reporting is careful to note these aren’t wheels or legs carrying it around the house. It draws on personal information such as emails to build a deeper understanding of its owner, and is meant to become more personalized and proactive over time, anticipating needs and surfacing information before being asked. Sam Altman has reportedly told staff it would be “the coolest piece of technology that the world will have ever seen,” and he and Jony Ive have previously described the device as “peaceful,” an “active participant” meant to bring users “joy” without being annoying.
The Jony Ive Talent Machine, and the Friction Inside It
The project sits inside OpenAI’s broader hardware bet built around Ive, Apple’s former chief design officer, whose startup io Products was folded into OpenAI last year in a deal valued around $6.4 to $6.5 billion. The endeavor has pulled in a remarkable concentration of ex-Apple talent: Chief Hardware Officer Tang Tan, who led iPhone and Apple Watch product design across 24 years at Apple; Evans Hankey, Apple’s former head of industrial design, now leading the speaker’s development; and Paul Meade, who ran Vision Pro hardware engineering before joining OpenAI last month. Ive’s design studio LoveFrom is driving the look and feel, with Ive reportedly making the final call on nearly every design decision. That arrangement has created internal tension worth watching: LoveFrom operates separately from OpenAI and hands designs over to OpenAI’s own engineers to build, and some employees have reportedly complained about the studio’s secrecy and slow pace on revisions, a culture clash between Apple-style deliberate iteration and OpenAI’s ship-fast reflexes that becomes a real execution risk once physical manufacturing is involved.
An Apple Lawsuit Looms Over the Timeline
The disclosure lands just four days after Apple sued OpenAI, io Products, Tan, and engineer Chang Liu in the Northern District of California, alleging systematic trade-secret theft. Apple’s complaint accuses Tan of using confidential Apple project codenames to interrogate job candidates, directing Apple employees interviewing at OpenAI to bring “actual parts” like batteries and logic boards for “show and tell” sessions, and circulating an Apple offboarding document to coach departing hires on evading exit-security checks. Liu is accused of downloading dozens of confidential hardware files and failing to return an Apple laptop. Apple says more than 400 of its former employees now work at OpenAI and calls the allegations the “tip of the iceberg.” OpenAI has denied wrongdoing, saying it has “no interest in other companies’ trade secrets” and isn’t aware of evidence the complaint has merit, and it maintains the screen-free speaker differs enough from anything Apple sells that it’s unlikely to implicate Apple’s IP. The stakes are concrete: Apple is seeking not just damages but an injunction against OpenAI’s hardware business, which could delay OpenAI’s ability to ship devices at all.
Insight: Why a Speaker, and Why Now
Starting with a stationary home speaker rather than a phone or glasses is a shrewd hedge. It’s the lowest-risk way for a software company with no manufacturing track record to learn hardware, supply chains, and retail before attempting anything worn on the body or carried everywhere. It also sidesteps direct confrontation with the smartphone, a market where Apple and Google would crush a first-time entrant, while still planting OpenAI in the home. The device is one of roughly five products in OpenAI’s pipeline, which reportedly includes an in-ear wearable codenamed Sweetpea, a possible pendant, smart glasses targeted for 2028 or later, and eventually a device meant to replace the smartphone. The speaker is the beachhead, not the destination. The timing also isn’t accidental: OpenAI is reportedly heading toward an IPO in the coming months, and a credible hardware roadmap plus a revenue stream beyond ChatGPT subscriptions strengthens the equity story, which is exactly why an Apple injunction threatening that roadmap is more than a legal nuisance.
Insight: The Privacy Problem Is the Product
The same capabilities that make this device novel make it a privacy lightning rod. A camera watching the room continuously, no wake word, ambient conversation capture, facial recognition, and email access aren’t bugs to be patched, they’re the core of the value proposition, which makes them impossible to quietly opt out of without gutting the product. Amazon and Google both took reputational damage when it emerged that human reviewers listened to Alexa and Google Assistant snippets, and those devices at least waited for a trigger phrase. OpenAI hasn’t publicly detailed how it will handle continuous audio, the persistent video stream, or biometric data, and it’s entering homes at a moment when trust in big tech’s handling of personal data is already thin. How OpenAI answers the “who sees the recording and where is it stored” question may matter more to adoption than any feature on the spec sheet.
Insight: This Redraws the Smart-Home Battle Lines
If it ships as described, the speaker doesn’t just enter the Echo-and-Nest market, it proposes a different category: an ambient AI presence rather than a reactive command box, a bet that frontier models can deliver something architecturally different from assistants whose basic design hasn’t changed since the Echo launched in 2014. The competitive response is already forming. Apple is preparing its own family of AI home devices, starting with a long-delayed command center codenamed J490 featuring a square 7-inch display, videoconferencing, and facial recognition, plus a version with a screen mounted on a repositioning robotic arm and a smart-home security system. Amazon and Google are both layering their own large language models onto existing speaker lines. The immediate market signal was narrow but telling: shares of Sonos fell more than 10% in late trading on the news before paring losses, while Apple dipped less than 1%. The read-through is that investors see the clearest near-term casualty not as the trillion-dollar incumbents but as the standalone audio-hardware players with no frontier model of their own to fall back on.
What to Watch Next
Three threads will determine whether this becomes a product or a footnote. First, the injunction: if Apple secures one, OpenAI’s 2027 timeline slips and the IPO narrative takes a hit. Second, execution: turning LoveFrom’s designs into a manufacturable device at scale is precisely the kind of problem that has humbled software companies before, and the internal friction over pace is an early warning sign. Third, trust: OpenAI has to convince mainstream buyers to put an always-watching, always-listening camera in their living rooms, something no company has yet pulled off at scale. The unveiling is targeted for later this year, which will be the first moment the pitch meets the public rather than anonymous sourcing, and the first real test of whether “a ChatGPT that lives in your home” is a category or a liability.
Leave a Reply