Interesting. Definitely contrary to what you would think. But The Heritage Foundation is a biased source and the numbers are pretty old. Could be interesting to see some new ones. Also the 75% identifying as whites in 2004 is in the entire population - not young people, which is the relevant demographic here.
How is the Heritage Foundation biased when it comes to military demographics?
As a Conservative organisation they are more likely to prefer a picture of the military as representative of the median population. Not saying they manipulated the numbers, but they do have a political agenda and are generally not an objective institution. As I mentioned the article does give a wrong impression of the ethnic makeup of the military by comparing the percentage of White soldiers (75%) with the percentage of Whites in the average population (73%) even though the white share of young people - which is the relevant demographic - is of course lower. So at least 1 manipulation.