OpenAI Is Building a Family of Devices for Its AI Models
OpenAI president Greg Brockman confirmed the company is developing a family of hardware devices designed for interacting with its AI models. No product details, timelines, or pricing have been disclosed.
Original sourceIn an interview with The Verge, OpenAI president Greg Brockman confirmed that the company is actively building a family of hardware devices intended to serve as dedicated interfaces for its AI models. The announcement is light on specifics — no form factors, release dates, or pricing were shared — but it signals OpenAI's intent to extend beyond software and into the physical layer of AI interaction.
The move comes in the context of OpenAI's acquisition of Jony Ive's design firm io in 2025, a deal widely understood as the foundation for a consumer hardware push. Brockman's confirmation of a 'family' of devices suggests the effort isn't a single product but a platform-level bet on new interaction paradigms — potentially spanning form factors from wearables to ambient computing devices.
The hardware market for AI-native devices has seen early attempts from companies like Humane (AI Pin) and Rabbit (R1), both of which struggled to find product-market fit. OpenAI enters with significant advantages: a massive existing user base, state-of-the-art models, and the design credibility that comes with Ive's involvement. Whether those advantages translate into compelling hardware remains to be seen when actual products ship.
For now, 'family of devices' remains a strategic signal, not a product announcement. The absence of demos, specs, or a release window means this is OpenAI staking a position, not delivering on one. The industry will be watching closely given how badly the first generation of AI hardware stumbled.
Panel Takes
The Skeptic
Reality Check
“Humane and Rabbit both announced ambitious AI hardware visions and shipped expensive paperweights — so forgive me for not upgrading my excitement level based on a Greg Brockman interview with zero product details. The specific scenario where this breaks: OpenAI ships a device that's essentially a ChatGPT app with a battery, charges $400 for it, and finds that people would rather just use their phone. The one thing that could make this real is if Jony Ive actually designed something that changes the interaction model, not just the enclosure.”
The Futurist
Big Picture
“The falsifiable thesis here is: by 2028, the dominant interface for AI will not be a keyboard and a chat window on a general-purpose device, but a purpose-built form factor optimized for ambient, voice-first, or embodied interaction. That bet requires two things to go right — AI capabilities becoming reliable enough that a single-purpose device doesn't feel crippled, and users developing enough AI-native habits that dedicated hardware feels like an upgrade rather than a regression. The second-order effect nobody is talking about: if OpenAI owns the hardware layer, it gains a distribution and data moat that makes its current API-dependent position look fragile by comparison.”
The Founder
Business & Market
“The buyer question is the only question right now, and nobody can answer it because there's no product. Consumer hardware is a brutal, capital-intensive business with razor-thin margins where even Apple took decades to build reliable distribution — OpenAI is walking in with model expertise and Ive's design reputation, but zero retail infrastructure, zero hardware supply chain credibility, and a category that just watched two well-funded entrants fail publicly. The moat they're building for is vertical integration: if your AI model, your operating system, and your device are all the same company, switching costs compound. That's the only version of this business that survives long-term, and it requires executing flawlessly on all three layers simultaneously.”
The PM
Product Strategy
“The job-to-be-done is still undefined, which is the entire problem with this announcement — 'interacting with AI models' is not a job, it's a capability, and until OpenAI tells us what specific friction they're removing from a specific user's life, this is a roadmap slide pretending to be a product strategy. The Humane AI Pin failed not because the hardware was bad but because nobody could articulate what job it was hired for better than 'your phone, but worse.' OpenAI needs to ship something opinionated enough to be wrong before it can be right.”