Consider this example:
https://www.cve.org/CVERecord?id=CVE-2026-113119
At this time, the corresponding Chromium bug tracker entry is still private:
https://issues.chromium.org/issues/557075191
(I just picked this one at random, I haven't checked if the information is publicly available from another source.)
For example: https://www.cve.org/CVERecord?id=CVE-2026-50000
Assigning the CVE happens much earlier than actually publishing, but it's the latter that ends the embargo.
What did happen is that a whole bunch of people had no clue and called the fact that "these 2 CVE IDs will be used for a PHP vulnerability" a CVE leak. Posts were going around yelling "CVE leak!" with a CVE database link that said "no information yet". Part of this is the mushed usage of "CVE" to just mean "vulnerability" rather than "CVE record". But still, quite ridiculous, and annoying. Information that CVE ID XYZ will be used for project ABC is *not* a leak. Like you say, CVE publishing matters. Getting the CVE record contents would be a leak.