Everyone expected a phone, but OpenAI’s first device is a ChatGPT-powered speaker that moves
OpenAI is building a portable, screenless speaker pitched internally as “a humanlike AI companion that lives in the home.” It’ll do the usual smart-speaker things—control your appliances, play music, answer questions, handle messages—all running on ChatGPT. The catch is what OpenAI wants it to feel like. The device incorporates mechanical parts that shift on their own, creating the sense that it’s alive rather than a box waiting to be told what to do, according to Bloomberg’s Mark Gurman.It’s still in development. OpenAI is aiming to reveal it later this year and ship it in 2027, as the first of roughly five devices in the pipeline—the phone many expected is further down that road, not the opening act. Internally, the company doesn’t even think of it as a speaker—it describes the product as a new kind of computer built for the AI era, one meant to make busy people more productive.
What OpenAI’s speaker actually does
The real sell is proactivity. Bloomberg says the speaker gets more personal the longer you own it, learning your habits and even drawing on your emails to figure out what you need. It’s meant to anticipate—surfacing information before you ask, playing the role of an assistant that already knows the answer.A camera and sensors let it read the room, literally, giving it context that current-gen smart speakers couldn’t be able to match. The voice comes from GPT-Live, OpenAI’s upgraded voice mode from earlier this month, which can listen and talk at once and adjust mid-conversation. And because it runs on a battery, you can carry it from the kitchen to the bedroom instead of leaving it tethered to one outlet.The muscle behind it is telling. OpenAI has former Apple design chief Jony Ive on the project through his studio LoveFrom, and Apple’s ex-industrial design head Evans Hankey is leading development of the speaker itself. Investors are watching closely too—Bloomberg noted Sonos shares tumbled more than 10% in late trading on the news.
OpenAI says its speaker is different from anything Apple sells, but Apple sells HomePod
That’s the crux of the fight. Apple sued OpenAI last week over alleged trade-secret theft, calling what it found “the tip of the iceberg.” OpenAI’s defense leans on distance—it argues the speaker “veers significantly” from anything Apple has on the market today. The problem is Apple already sells the HomePod and HomePod mini. OpenAI’s answer is that the audio system and underlying hardware are substantially different, and that it doesn’t see Apple’s speakers as comparable to what it’s building.Proving that may be the easy part. The harder issue is people. OpenAI spent $6.5 billion last year to buy io Products, the hardware startup co-founded by former Apple design chief Jony Ive, and has since hired more than 400 people from Apple—the recruiting spree Apple’s lawsuit is really about. OpenAI denies wrongdoing, saying it’s seen no evidence the complaint has merit. But Apple wants an injunction that could stall the launch, which means 2027 is a target, not a guarantee.There’s irony in the matchup, too. Apple is quietly building its own family of home devices, including a tabletop robot with a display on a moving arm—not so far from the animated, “alive” machine OpenAI is now racing to ship first.
