It really depends on how willing any given court is to ignore obvious facts and unresolvable contradictions.
Harvard argues in one part of their brief that “race conscious” admissions are good because it produces “racially diverse” outcomes. In another they deny that race is considered at all and assert that they just happen to keep changing their admissions criteria every few years in ways that just by complete coincidence manage to arrest any rise in Asian admissions rates. It can’t be the case that Harvard is both engineering racial diversity and not taking into account race, but a favorable court will ignore this whereas a court less inclined to go down the path of “anything labeled affirmative action is good and anything we think is good is legal” might not ignore this inconsistency.
It’s also known to everyone involved (Harvard, its supporters, its opponents, the judges) that there is in fact a simple numerical adjustment to the admissions score for all non-white applicants, in which blacks receive a bonus and Asians receive a penalty. Harvard calls this the “personal rating” and claims they are evaluating “humor” and “grit,” which it claims with a straight face are relevant metrics for university admissions, but it correlates exactly with race. No one actually doubts that this practice goes on, but supporters of the Harvard policy have to prevaricate and pretend that the “personal rating” either does not exist or is rating something other than race.
I think it is a social good to stop rewarding people who engage in collective bullshitting and the people defending the Harvard policy should be, at a minimum, forced to be honest about what they support and why before a court will consider endorsing it.