Yes, with HRM the ascospores can survive pasteurisation and germinate following the thermal process.
Control of these is really about the raw material as you noted - buying the right product to the right specification, ensuring that your raw material suppliers have suitable controls etc.
If the level of contamination in the raw material is low, you may not get a particularly even distribution in the production - as these are physically "solid" they wouldn't necessarily have a homogeneous distribution even after mixing, as you'd expect to see with a liquid/soluble contaminant. Nonetheless it would be a little unusual. I'd still undertake some identification work on the mould anyway, as even if it isn't HRM it may be useful knowing what it is so you can potentially link it to other possible sources.
It could also be the case that only some of the bottles provide exactly the "right" environment to permit growth - it's not something I recall seeing for mould, but the similarly challenging Alicyclobacillus can do some unusual things like this, e.g. only grow in some packs where the fill vol/headspace is just right.
If this is a new aseptic process you've set up then I think I'd still be focussing my attention there - these can be extremely sensitive and fiddly to get running correctly. For example I have seen sporadic micro problems (similar to yours - some bottles but not all) caused by problems with the assembly of the filler heads, but this was only identified after completely disassembling and inspecting under magnification - normal visual inspection suggested it was fine, but it suffered a tiny crack during assembly and compromised the aseptic status enough to cause problems.