During a press conference in San Francisco where Qualcomm's senior director of wearables, Pankaj Kedia, revealed the chip, it was apparent that the 3100 is all about one thing: battery life. In fact, there was barely a mention of any other improvements we can expect, beyond a better ambient display.
Essential reading: Best Wear smartwatches to buy
With the 3100, Qualcomm is putting battery life first – which is exactly what smartwatch makers have been asking for. Will there be general performance improvements like a faster UI? Qualcomm wasn’t very forthcoming with answers when I asked, and it's probably because there aren't any. And it's not all that surprising.
To explain why, we need to first look at how the new chip is different. The headline act is undeniably Qualcomm’s new co-processor, which is going to take on a lot of the workload when you’re not actually tapping on your smartwatch. Qualcomm says manufacturers will be able to utilise the co-processor for various background tasks, like heart rate or sleep tracking, which operates at 20x less power than the main processor.
Sounds great, right? But then you see that main processor is the same one that was on the 2100, an ARM quad-core A7 CPU cluster. The difference this time is that those cores are being used more efficiently – great – but that processor is now more than two years old. So while we're hearing about 7nm processors in the smartphone world, why is the smartwatch still so behind?
The 3100 feels more like a workaround than a leap forward
Getting a smaller process on a chip would require an entire re-engineering of the architecture by Qualcomm; it’s not just a case of taking the 28nm process and firing a shrink ray at it. And guess what? New architecture costs a lot of money. Smartwatches are a drop in the ocean of Qualcomm's smartphone business, and until they’re selling at significantly higher higher volumes, it’s unlikely Qualcomm is going to put in the R&D to make something smaller.
Interestingly, Qualcomm told Android Police that the decision not to shrink the 28nm process was because the passing drain on a 10nm process is actually more than a 28nm, and because a smartwatch is idle most the time, it made sense to keep the same size. That's an interesting claim, but I'm not wholly convinced that it's a better solution than a complete re-working of the architecture. It's also certainly only beneficial for battery life.
But all of this is why the 3100 feels more like a workaround than a leap forward. OEMs want longer battery life above anything else, and with the new chip they're getting it. Qualcomm will see a return here, and while improved speeds and other performance improvements would be nice – and would certainly come from a 10nm process – they're not realistic right now. I'd love to see Qualcomm throw caution to the wind and double down on smartwatches; if I ran the business I would probably feel differently.
So I come away from Qualcomm's news with mixed feelings. It's probably not going to galvanise Wear OS in the way many, including myself, had hoped, but it's going to keep things ticking along.