This isn't just on Pixels, I have a 6P that will turn off early as well, and I'd hazard a guess it has more to do with the latest Android versions than the specific hardware.
Literally fluctuates between 40%, 1%, 31%, off and don't get me started on how it changes when it gets cold outside.
https://www.apple.com/uk/support/iphone6s-unexpectedshutdown...
To prepare your iPhone 6s for the battery replacement process, please follow the steps below:
Back up your data to iTunes or iCloud.
Turn off Find my iPhone
Erase data and settings in Settings > General > Reset > Erase all Content and Settings
But this is just the nature of batteries.
Anyways, I did what most people do not do and unlocked bootloader, installed TWRP and Pure Nexus Project 7.1.1 ROM and am happy to report that issue is now completely gone (I did stress-test it several times in below zero temperatures) !
So, no - it does not seem to be a hardware issue.
And as an aside, I dropped my Note 4 recently and broke the glass. I was thinking about the possibility of getting a new phone. But after looking at this season's options I'm heavily leaning toward just replacing the glass. One reason is that no top-of-the-line smartphone seems to have a replaceable battery to avoid these problems.
More generally, this season's phones are also cheap-looking and cheap-feeling. The Pixel feels like a $200 phone at best, with its cheap aluminum shell, fragile-feeling uncomfortable buttons, and ugly plastic/glass/whatever square on the back. I've been losing faith in Samsung's build quality (I've had some issues with my Note 4 and my previous Samsung phone, and the news about the exploding Note really turned me off the company). Even the latest iphone looks uglier than usual (that glossy bezel really doesn't do it for me, looks like a free phone you get when you sign up for a phone plan).
2016 is a really really bad year for smartphones. None of these machines seem like they're really worth their price. The next great phone is probably going to be one that doesn't feel like corners were cut.
Now, why this happens in a brand new phone like the Pixel is another matter...
I went from a Note 4 to a Note 5, and ended up dropping my Note 5 causing the screen/LCD to break and not turn on. It looks like it costs about $300-400 to replace everything after asking a couple of local places, because the parts for this are expensive/out of stock. You might want to see if it's the same thing for the Note 4.
At that price, I'd probably just want to buy a new one.. but as you said, there's not a lot of good smartphones out right now.
I'm using the Note 4 again for now, which is actually a refurb/replacement because the original one's screen died randomly one day (without dropping it this time.)
One thing that really bothers me about a lot of the new phones is the slippery glass backs. The Note 4 has a textured plastic back with quite a bit of grip to it. I tried a friend's Galaxy S6 and it nearly slid out of my hand!
I suppose a silicone case would fix that, but I like being able to use the Note 4 without a case and not have to worry that it will slip away.
If it is software, I'd strongly suspect something at kernel level, since quite a few times in my experience, the shutdown is too quick for a safe shutdown. Userspace seems to get a notification that battery is low because it usually gives the "shutting down" alert, but then just shuts off in a suspiciously short amount of time.
I'd be more likely to think it's probably something in the battery/power stack that's not managing or reporting the power well. Discharge/voltage curves are not at all linear, [1] [2], so it's logical that issues with immediately available voltage might start to affect the device around the 30% level. Additionally, as another commenter has mentioned, the discharge curve is dependent upon temperature [3], so a sudden drop in ambient temperature will eventually have a direct effect on the reported voltage.
I'm not exactly sure what kind of metrics are commonly used for reporting battery percentage, but I am about to read a few (okay, maybe half of one) Arxiv papers that came up in a search. [4] Presumably, it's strongly based on some voltage modeling and I can imagine that would have a hard time coping with external factors beyond its measurement capacity, like ambient temperature drops. I don't suppose a user would be happy to see their reported battery percentage drop like a rock immediately upon stepping outside.
I really don't know where I was going with all that, except I guess to say that my gut says Battery "percentage" is a bit more an art than a science. That is to say, there are psychological reasons not to report based on voltage directly (it may go up if you step into a warmer room; "Wat?"), and I can imagine any metric is going to have its strengths and weaknesses, because you really can't account as much for user behavior that changes important physical conditions of the battery.
[1] http://www.powerstream.com/AA-tests.htm
[2] http://www.ibt-power.com/Battery_packs/Li_Ion/Lithium_ion_te...
[3] http://www.ibt-power.com/Battery_packs/Li_Ion/Li_Ion_DiscTGp...
[4] https://www.google.com/search?q=battery%20percent%20site%3Aa...