Qualcomm’s latest chip could lead a new wave of camera-equipped AI watches and wearables
I’ve been bracing myself for the coming wave of AI-powered wearables that can be worn everywhere, based on reports from gadget systems on. Meta, Google again an apple — a halo of connected technology with cameras on board, streaming in AI services. Qualcomm’s latest chip, announced Monday at the Mobile World Congress in Barcelona, is designed for it, and the first devices using it are coming this summer. Samsung, Google and Motorola they already build the hardware with it.
I sat down with John Kehrli, Qualcomm’s senior director of product management, to discuss the new wearable chip push, and it caught my attention on several levels. The reason you care is that this is a clear preview of upcoming tech products: Qualcomm’s power chips are almost all non-Apple ones. watches, VR headsets again smart glasses out there.
While Qualcomm has had separate chip lines for smart watches and smart glasses and VR headsets, the new Snapdragon Wear Elite chip aims to unify the categories. It’s a powerful watch full of different capabilities for wireless communication, but it’s also designed to support video input and AI streaming, even 1080p video output on displays. That could include AI-infused smart glasses.
“It’s not just the watch: that’s certainly our focus, but the portfolio [of devices] it’s grown a lot,” Kehrli said.
Here are the stories about the Snapdragon Wear Elite that stood out to me.
Qualcomm’s new chip design is meant to be flexible in nature. It may end up in many places.
Multiple in-line AI offline processing
A big part of Qualcomm’s push for these chips is to do productive AI and LLM work on the device, a trend I expect to grow. The Snapdragon Wear Elite looks more powerful than previous Qualcomm watch chips. Some of the offline, on-device functions could be voice-based AI, for fitness or, according to Qualcomm, “for life hacking.”
I’m not sure I need life hacking, but I would be interested in having more AI-based controls for wearables. Additional capabilities look like driving video to displays and using internal cameras, including video streaming. The whole idea behind next-wave multimodal AI is for AI services to be aware of what you’re doing — that will happen mostly with camera access.
Kehrli says that the processing cores of the neural processing unit in the Snapdragon Wear Elite can support AI models of up to two billion parameters on the device, about 10 tokens per second for processing. He sees that as good enough for most offline needs, with cloud-connected AI kicking in when needed.
Kehrli sees many local AI needs for additional sensors, including cameras, that will be in these wearables. “There are a lot of interesting ideas coming up [to the devices]. Location, sound, voice, text, all sensors — we’re really seeing more medical-grade sensors coming to the retail space. What do I do with that data?”
Qualcomm’s concept for a wearable pendant is like a smart watch, but with an outward-facing camera.
Cameras everywhere?
In Qualcomm’s sizzle video of the new chip, we can see a glimpse of a watch with a camera on its top edge. Most smartwatches don’t have cameras right now, but that may change soon. Although not a good way to take pictures, internal cameras are probably an additional way to get into AI, such as facial recognition biometrics to tap to pay, using a watch like a smart key for cars or other connected devices, or perhaps to use for other AI-based controls.
Another image of the pendant, which looks like a smartwatch worn around the neck, has a camera facing outwards. All of the AI pins and pendants over the past few months have featured similar ideas. Like smart glasses, outdoor cameras can be an alternative way to see things without putting something on your face. But you will have to wear a pin or pendant.
Qualcomm talks about the new chipset.
Better battery life, fast charging, low power Wi-Fi connection
And it sounds like devices with these new chips will last longer on a charge. Qualcomm’s promising 30% better battery life than its previous clock chip — potentially “days” of use. I would have expected more or less a day, considering these chips may charge more for camera-based features and AI.
Fast charging sounds promising, though. The chips can charge devices up to 50% in 10 minutes of charging. That’s key because many of these wearables are designed to be worn all the time, and some while you sleep. It’s like companies are trying to find ways to create a quick recharge pit without spending too much time away from your body.
The most interesting part may be the advanced wireless features. Qualcomm has six different protocols on the chip: support for Redcap 5G (a protocol to support high-speed and low-power connected technologies), Bluetooth 6.0, ultra wideband, GPS, NB-NTN satellite communication, and micropower Wi-Fi 802.11ax.
Low-power Wi-Fi support could allow these new wearables to stay connected to Wi-Fi continuously, Kehrli said, allowing them to work in the background for longer. Meta’s Ray-Ban glasses, for example, currently mostly connect via Bluetooth and don’t stream video automatically; switching to that mode kills battery life quickly. Streaming AI modes can take longer on more powerful Elite devices.
Qualcomm’s plans for this chip extend to almost every wearable area.
Where they can appear: Watches, glasses, headphones, pendants, and more
Qualcomm aims to put its new chip in a variety of wearables, from camera-enabled headphones and earbuds like Razer’s Motoko concept (which I tried at CES in January) to next-generation smartwatches and AI paints, to smart glasses, and even sensor-connected bands. Similar devices Meta’s neural band, that uses EMG (electromyography, which uses skin touch sensors) with hand gestures to control its smart glasses, can see improvements with this chip. Maybe that’s exactly the kind of place Meta could explore report smartwatch debut this year.
It’s also clear that everyone, including Qualcomm, isn’t entirely convinced that people will choose to wear these future AI gadgets. Are the glasses? A pendant? Watch? Headphones? All of the above? Kehrli feels that people will have different options and will choose what works. Will that kind of redundancy make sense or settle into clear categories in another year or two?
Glasses, Kehrli adds, could be the destination for this chip because of the possibilities of connecting mobile devices, saying that he expects the adoption of wearables with their data connection to continue to increase, especially with AI services. “We see, manually, up to 50% of customers are connected [wearables] by service plan. We’re seeing that grow a lot, especially with this AI in the cloud-locked device type.”
It’s clear that the halos of wearables are on the floor of a few big companies. How it all goes and works, however, is still unclear. And while these new wearables should be very powerful, currently the focus is not on improving how they can stay connected and communicate with each other, something I saw little of in the personal mesh network demo. done by starting Ixana at CES. Maybe that’s what’s next on deck.
Meanwhile, wearables are trying to be better extensions of your phone, first, and perform better as standalone devices as well.
