Then you expect wrong, at least about Database Management Systems. New performant ones are often (/ mostly?) written in C++. At least that's what it's like for analytical DBMSes.
Some examples:
* HyperDB: https://www.hyper-db.de/
* DuckDB: https://duckdb.org/
* Vectorwise/Actian Vector: https://www.actian.com/analytic-database/vector-analytic-dat...
And you might also be surprised that many problems originally carried over from C to C++, like dereferencing nulls and memory leaks can be done away with - not by being super-careful, but by sticking to newer language facilities and using some static analysis. Other, like buffer overruns, are easier to avoid with things like spans and ranged-for loops.
Why even bother stating this? “I don’t have any data to back this up but look at this huge number!!!!”
I know the $ amount is greater than all other security risks in the history of software - even if I don't know the number the only reason to NOT post something about that history would be to defend C/C++, when in fact I am attacking C/C++'s record here.
I get the reasoning to learn it. Just after you do learn it, stop and use something else. Unless you absolutely have to. I don't think anyone has attempted to compute the amount of money spent on buffer overflow hacks. But I'm guessing it's in the 10s to 100s of billions of dollars if you count all the viruses that took advantage of it in the past in Apache web server or Windows NT's NTFS (or sharing etc...) codebases. And that applies to C++ too.
And hey, I know that sounds extreme - but there really are situations of "unless you absolutely have to" but I wouldn't expect anyone to write a brand new queuing system or database in C or C++ anymore given the alternative languages available.