As of late March, it is fairly clear that only a small fraction of US COVID-19 cases are being identified through testing. (Comparing for example the fatality, confirmed case, and CFR numbers shows this.)
Tests are ramping up (albeit slowly) in the US but current strategy seems to only aim to test symptomatic cases or high-risk groups.
It is possible, though, that the only bridge measure to a vaccine that allows less than tens of percent of the US to be infected with COVID-19 is blanket testing and followup; this would lead to a huge number, potentially hundreds of millions or even billions with repeated testing.
Resolution will be based on credible and corroborated estimate of the total number of tests administered by June 1, +/- one week. Multiple credible estimates as judged by moderators can be averaged.
Concerns tests administered, not people tested. (And multiple tests used at one time to increase accuracy count as more than one.)
Any form of test that involves a physical specimen and some chemical process counts (but not say an app that takes your temperature and listens to you cough.)