Bloomberg’s latest report on OpenAI’s first consumer device generated the familiar wave of excitement:
Jony Ive, Sam Altman, a mysterious screenless AI companion and a hockey-puck-sized object that may cost $300–400.
But strip away the visual drama and the new information is relatively thin. The puck-like design circulating online is an old, unofficial concept render, not a confirmed OpenAI product image and the core story has been known since OpenAI announced its $6.4bn acquisition of Ive’s io startup in May 2025.
What the latest reporting really adds is a prospective price range, a more specific product category, that is portable, screenless AI speaker, and a suggested 2027 launch window. That makes it more tangible, but it does not yet establish a compelling new computing category.
The real news is Apple
The more consequential development is not the alleged design or the $300 price point. It is Apple’s trade-secrets case against OpenAI, io and former Apple employees. We did cover this couple of weeks’ ago from another perspective.
Apple filed suit in July 2026, alleging that confidential information related to its hardware development was improperly taken to OpenAI. In early August it escalated by seeking a preliminary injunction: an immediate court order that could bar the defendants from accessing, using or disclosing the alleged secrets while the litigation is heard. Apple has also sought expedited discovery.
OpenAI has denied the allegations and moved to dismiss the case, arguing that Apple has not identified specific trade secrets with sufficient precision and that OpenAI does not possess or require them.
For now, that means the product programme can continue.
But the legal proceedings run in parallel with development, creating a material execution risk. A court-granted injunction would not necessarily kill the device, but it could restrict staff, evidence, suppliers, design decisions or certain areas of development, and potentially force delay or redesign.
What is the puck meant to do?
The reported product is best understood not as a new type of silicon or a radical hardware invention, but as an attempt to build an AI-native ambient interface.
Rather than opening ChatGPT on a phone or laptop, a user would speak to an always-nearby device. Its proposed value is persistent, multimodal context: voice interaction, potentially environmental sensing and camera input, memory of ongoing tasks, and access to ChatGPT’s models and agents. Bloomberg’s earlier reporting described a mobile screenless speaker intended to function as an AI companion, possibly using movement to feel more responsive or “alive.”
That creates a simple pitch:
Your phone is where you open an app. The puck is where AI is already present.
In theory, it could handle lightweight planning, messages, reminders, questions, household tasks, media, contextual assistance and agentic workflows without requiring a screen. In practice, that promise depends on model reliability, latency, privacy controls, battery life, audio quality and whether users genuinely want a microphone, and perhaps a camera running in their living space.
The missing commercial detail
The headlines focus on $300–400 as though that is the price of an AI companion. It is almost certainly only the price of the object.
A hardware purchase will pay for the enclosure, microphones, speaker, battery, sensors, connectivity, local processing and margin. It will not sustainably fund unlimited cloud inference for an always-available assistant using frontier models, speech processing, memory, vision and agent/tool calls. The reports say the device will integrate with ChatGPT, which strongly suggests that a cloud account and recurring service economics sit behind the hardware.
The likely model is therefore familiar:
OpenAI has not announced pricing, subscriptions or usage terms, so this remains inference rather than confirmed policy. Still, the economics are difficult to avoid: the puck would be a route into a continuing AI service relationship, not a fully paid-up intelligent appliance.
We have seen this story before
The underlying idea of an always-available, voice-first, context-aware computer beyond the phone is not new.
Amazon, Google and Apple have already placed voice assistants in homes through Echo, Nest and HomePod. More recent startups tried to make generative AI into a dedicated object: Humane’s AI Pin and Rabbit’s R1 are the obvious examples. They attracted attention because they promised a break from screens; they struggled because phones remained faster, more reliable, more private-feeling and more useful.
The common failure mode is straightforward:
a dedicated AI device must beat the phone at a task users perform often enough to justify charging, carrying, updating and trusting another object. A voice assistant that merely calls the same cloud model feels redundant. A camera-equipped assistant that watches or listens continuously creates a far higher privacy threshold.
OpenAI and Ive do possess advantages those startups lacked:
ChatGPT already has enormous consumer awareness and an existing user base.
OpenAI has access to frontier models and can improve the service continuously.
Ive’s team can plausibly deliver better industrial design, interaction design and emotional appeal than a typical AI hardware startup.
A $6.4bn acquisition signals the capital and organizational commitment to persist beyond a first-generation product.
But those advantages do not eliminate the core product question: what can this do better than ChatGPT Voice in my phone, earbuds or laptop?
The puck’s possible future
There are three plausible futures for OpenAI’s device.
First, the subscription accessory. It becomes a beautifully designed premium smart speaker for committed ChatGPT users. It succeeds modestly as a home/desk endpoint, but never displaces the smartphone. This is probably the most commercially realistic near-term outcome.
Second, the agent interface. If OpenAI delivers genuinely dependable agents, deep contextual memory and privacy-preserving local sensing, the puck could become the preferred way to delegate routine tasks. In that world, the product matters less as a speaker and more as a persistent gateway to an AI operating layer.
Third, another AI gadget cautionary tale. If it has vague use cases, unreliable agents, high subscription costs, intrusive sensing or weak integration with the phone ecosystem, consumers may treat it as an expensive novelty. It then becomes evidence that “screenless” is an aspiration, not a product category.
For now, Bloomberg’s $300 headline is useful because it anchors an abstract project in a probable consumer price bracket. But it does not transform what we know. The puck’s shape is not the innovation, and neither is the price. The actual bet is that OpenAI can turn a cloud AI service into a trusted, habitual physical presence and do so while its hardware programme is under legal scrutiny from Apple.
Latest on Intelligent founder and Quantopinion.
