The authors assumed (as far as I can tell) that circumcision had an effect only on heterosexual transmission. This is a conservative assumption, so we should bear in mind that the results will also be conservative estimates.
They found that neonatal circumcision "reduced the 1.87% lifetime risk of HIV among all males by about 16%" (that is, it reduced by about 0.3% to about 1.6%). This effect varied by race, from 7.9% for white men to 20.9% for black men. They also found "lower expected HIV-related treatment costs and a slight increase in [Quality-adjusted life years]".
The CDC's forthcoming recommendations re neonatal circumcision are in the news at the moment (see, for example, Debate on circumcision heightened as CDC evaluates surgery [Washington Post, 19 Jan]), and of course there is much speculation about what those recommendations will be. To my mind, there's an interesting clue in the conclusions of this paper:
Racial and ethnic groups who could benefit the most from circumcision may have least access to it due to insurance coverage and state Medicaid policies, and these financial barriers should be addressed. More data on the long-term protective effect of circumcision on heterosexual males as well as on its efficacy in preventing HIV among MSM would be useful.