A good dictionary (which I assume Merriam-Webster is, I only know about the OED) meticulously collects samples to show that they're reporting the actual language as it's used, not some fantasy†.

So, either Merriam-Webster has some examples where "CubeSat" means these rather bulky ordinary satellites or, as seems more likely, if they re-examine the material they'll realise it's a mistake and the definition must be corrected.

In the first era of dictionaries this was painstaking work, you could expect that if you discover a mistake, you write a letter, a week later some expert reads it, and maybe a year later there's a correct draft, but it might take a decade until the dictionary publishes errata. But today there's a fair chance if you can contact Merriam-Webster about their mistake they can publish a revision (to their online dictionary) the same week.

† This choice leads to several fun stories, e.g. https://en.wikipedia.org/wiki/Twat

Yes, all dictionary makers have large files showing examples of current usages.

The OED is different in that (a) it retains historical examples rather than just current ones, and (b) it publishes that history.

You don't use the OED the same way as a conventional dictionary. Conventional dictionaries are used to check whether you're using a word right, or to understand a word you read in a contemporary document. The OED is for historical research, to study etymology or understand old works.

In this case, it's probable that they just screwed up the math. But they do have a sample they can point to.

I think it's a conversion error on their part, 1000cm3 != 1m3

Agreed, the MW definition does not seem to be correct.