AI wearables: Hype, wreckage, and one part that’s actually working

Two years ago, AI wearables looked like the next big platform shift. Pin it to your shirt, talk to it, never touch a screen again. That version of the story has mostly ended in failure. But a quieter, narrower version of it is genuinely taking off. Worth separating the two before deciding what to make of any of it.

What actually happened

Start with the wreckage. The Humane AI Pin was a small square badge that clipped magnetically to your clothing. It had no screen. Instead, it projected a basic display onto the palm of your hand and you spoke to it, asking questions, taking photos, sending messages, all through voice and a built-in camera. Humane launched it in April 2024 at $700 with a subscription attached. It sold fewer than 10,000 units, and by February 2025 the company had shut down the Pin’s servers entirely and sold its assets to HP for $116 million, a steep drop from the roughly $850 million it had been valued at not long before, according to GSMArena.

Rabbit’s R1 told a similar story from a different angle. It wasn’t worn, it was a small handheld orange gadget with a screen, a scroll wheel and a camera, built around what Rabbit called a “large action model” that was meant to carry out tasks on your behalf, like booking a taxi or ordering food, just by asking it. It sold around 100,000 units at $199, no subscription. But the company’s own founder admitted to Fast Company, in reporting picked up by Tech Times, that only about 5,000 of those buyers were using it daily. People bought the gadget, tried it, and largely put it in a drawer.

Both devices tried to do the same thing: replace the smartphone with a single, general-purpose AI assistant you wear or carry. Both discovered that “talk to an AI about anything” isn’t, on its own, a reason to buy new hardware.

What’s actually working

Compare that to what’s happening elsewhere in wearables right now.

Meta’s Ray-Ban and Oakley smart glasses look like normal glasses. They have a camera, speakers and a microphone built into the frame, and let you take photos and videos, listen to music, and ask Meta’s AI assistant questions hands-free, all with no screen on most models. Meta and EssilorLuxottica sold more than 7 million of them in 2025 alone, more than triple the year before, according to EssilorLuxottica’s own Q4 2025 earnings report as covered by UploadVR. That’s not a novelty number. It’s the kind of volume that makes a hardware category real.

Health wearables are doing something similar. Whoop is a screenless wristband that tracks sleep, heart rate and physical strain and turns that data into daily recovery guidance through its app, sold on a subscription rather than a one-off purchase. It now has more than 2.5 million members, grew bookings 103% year on year in 2025, and just raised $575 million at a $10.1 billion valuation, according to TechCrunch. Oura works on the same principle but as a ring rather than a band, tracking sleep and readiness. By its own count, Oura had sold 5.5 million rings and expected to double revenue to around $1 billion in 2025, and it filed confidentially for an IPO in May 2026, though as a self-reported figure it’s worth treating as company-stated rather than independently audited.

Why the split

The pattern isn’t “wearables work” or “wearables don’t.” It’s that AI wearables succeed when they do one thing well and fail when they try to replace the phone.

Smart glasses succeeded by being glasses first. You already wanted a pair. The AI is an addition, not the whole pitch. Whoop and Oura succeeded the same way, by being excellent at one specific job, tracking sleep and recovery, and layering AI-driven insight on top of data people already wanted. Humane and Rabbit tried to sell the AI itself as the product, with no existing reason for someone to put the object on their body or in their pocket. That’s a much harder sell, and so far it hasn’t worked.

Is this the future?

Probably, but not the version that got the headlines two years ago. The near-term future looks like AI quietly built into objects people already wear or use for a specific job, glasses, rings, bands, rather than a new standalone gadget competing with the phone in your pocket. The general-purpose AI wearable may still arrive eventually, but the current evidence points to specific, purpose-built devices winning, not universal ones.

The takeaway for Jersey businesses

If you’re evaluating whether AI wearables matter for your business, the useful question isn’t “should we get into wearables.” It’s narrower: is there one specific, recurring task our people or customers already do, that a wearable could make faster or more accurate. That’s the pattern behind every wearable success so far. The gadgets that tried to do everything are the ones sitting in a drawer.