I wonder if ESP8266 or similiar (ESP32 for example) modules are used often in real mass production in medical/automotive/military/households industries.
I see many DIY projects with it, but there is a better alternative for mass production?
If yes - i need some examples.
The ESP8266 chip is relatively new to the market (circa 2014, I think). So it's short market availability has limited its adoption somewhat -- some early adopters have chosen it, but that's usually not a benefit to high volume, low cost designs. There are also many more well tested chips/modules that have better documentation and better support. TI, Silicon Labs, and Microchip all offer chips with similar features (each manufacturer even offers several different versions targeting specific design constraints).
One of the major selling features of the ESP8266 is its low cost (a huge benefit for the maker/hobbyist market). But when buying any of the other existing parts in large quantities (for mass production), the prices become pretty similar.
A quick search shows a few products that use the ESP8266 chip. There are likely many more products that use the ESP8266 as well, but their bill of materials is not publicly available. Here's two that came up immediately:
Sonoff Smart Switch, Ebay Light Controller
It's likely that the ESP8266 (and ESP32) will eventually be adopted by more and more designs, especially as the "Internet of Things" industry grows and engineers become more comfortable using them.
To answer your specific question, 'is there a better alternative for mass production', of course there is. That alternative is the same for any module, nothing I am about to say is specific to the ESP8266. The better alternative to using the ESP8266 module in the context of mass production is using the ESP8266 chip. The modules have the advantage of already being compliance tested, which makes dealing with regulatory agencies, such as the FCC, much much easier, and not having to deal with antennas or other RF design considerations.
This comes at the disadvantage of higher cost, a bulky form factor with low integration into your product (though the SMD modules do a fair job of mitigating a lot of this), and little flexibility on the RF side of things. You're antenna, as well as range etc. are at the mercy of the modules available.
When the mass production is mass enough that the cost savings of using the chip and your own support components over the module, spread over the entire production run, justifies the extra development cost and getting FCC approval etc, then it is a no brainer: don't use the module.
Here it is on Mouser. Check how many are in stock.
As for whether the ESP8266 is a good fit or not, that is up to you to figure out. Read the datasheet, compare similar chips, factor in both the BOM cost as well as the engineering cost (engineers are expensive) and you can fairly easily figure out if the ESP8266 is a good fit for your application or not.
If it is or isn't used in DIY projects is completely irrelevant to any of this. That carries with it no implication of if something suitable for 'real' products or not. The actual device, chip, module, or whatever itself determines that. But the ESP8266 is just another chip available from all the big distributors, and can be bought directly from EspressIf in even larger volumes. The modules can also be found from every distributor, though often in lower volumes (thousands vs. tens of thousands) and again, directly from EspressIf. But that simply reflects demand. Higher volumes almost always use the chip directly and not the module, just because the numbers work out that way.
Should you use mystery modules bought on ebay in a mass production product? Of course not. But you can get the modules and the chip from any trusted distributor, exactly like every other part on your entire BOM, and being popular or unpopular in the DIY community has no relevance on any of this. If it is used by the DIY community, it usually just means the barrier to entry for development is low, both in the cost of hardware/software, and the documentation and tools available. In the case of the ESP8266, it's popularity is mostly due to the cost in low quantities.
I honestly can't think of a less useful metric determining parts for a mass produced commercial product.
Oh, and of course they're used in products, or distributors wouldn't be stocking them.
ESP8266 is just for learning . It should never be used in industries. It doesn't even meet the industrial standards.
The chip is very sensitive to fluctuating voltage levels or noises. It gets reset frequently due to stray signals. It's transceiver's signal strength is also very very less.
Most importantly, a high school kid having a little bit of experience in JavaScript and HTML can easily hack it. And obviously we don't want our industries to be shut down by crazy hackers
Industry tend to use circuits made specifially for their use. It winds up being cheaper that way, and there's no dealing with a third party when things go wrong or supplies run out.
Also, there's question as to whether the ESP8266 is RF compliant. The earliest device, the ESP-01, had no shielding at all, and others are known to have fake regulatory compliance logos (such as FCC).
