I don't think Andrew believes Zig is going kill C or C++, he probably has hope but I think he is aware of the reality. He found a way to make a living on something he was passionate about.
Use-after-free is a fact of life until something kills C, but the realities of language adoption are against that. Zig seems interesting and worthwhile in offering a different perspective on the problem and does it in a way more agreeable than Rust or the like for all those who love C and are adverse to large complex languages. The realities of language adoption are as much for as against Zig, large numbers of people are still getting drawn to C and Zig seems to do a better job addressing why so many are drawn to it than the alternatives.
For that to matter OS vendors that only care about C on their platforms, have to also care about Zig.
Otherwise the only users are going to be the ones happy to do some yak shaving instead of the actual application code with the vendor tools.
It also ignores that C doesn't stand still, the competition is C2y, not C89.