And as they note, these are hand-wavy numbers and estimates.
Note. These numbers do not represent actual cases in persons with a diagnosis of HIV infection or AIDS. Rather, these numbers are point estimates of cases diagnosed that have been adjusted for reporting delays and for redistribution of cases in persons initially reported without an identified risk. The estimates have not been adjusted for incomplete reporting.
Data include persons in whom AIDS has developed and persons whose first diagnosis of HIV infection and the diagnosis of AIDS were made at the same time.
"Over time, HIV infection may progress to AIDS and be reported to surveillance. Persons with HIV infection (not AIDS) who are later reported as having AIDS are deleted from the HIV infection (not AIDS) tables and added to the AIDS tables. Persons with HIV infection may be tested at any point on the clinical spectrum of disease; therefore, the time between diagnosis of HIV infection and diagnosis of AIDS differs. In addition, because surveillance practices differ, the reporting and updating of persons clinical and vital status differs among states. Completeness of reporting for HIV infection (not AIDS) is estimated at more than 85% [4]. By matching data in the national surveillance database, CDC estimates that approximately 2% of cases of HIV infection (not AIDS) are duplicates."
The people on these tables who have AIDS, by definition of the data, had HIV first, as reported! (With a few exceptions noted; ~20 cases of AIDS without any detectable HIV Ab). You do the math.