There is some distribution of effort/expertise at least:
When an individual researcher or an organization discovers a new bug in some product, a CVE program partner — there are currently a few hundred across 40 countries — is asked to assess the vulnerability report and assign a unique CVE identifier for the flaw if and as necessary.
barsoap@lemm.ee 3 days ago
That’s long since been the case, e.g. the Linux Kernel assigns its own CVE numbers, they’re a CNA. Which keeps the “root” CVS database completely out of the loop short of saying “this here is your namespace and scope”. Canonical is a CNA, Airbus is a CNA, both covering their own products. 453 in total.
Still important to have a fallback though because not all projects are big enough to do that kind of stuff.