Are you planning to produce infantry in squads, like we do, or perhaps produce them in squads and have them controlled like squads?
Either way, you might consider how different powers in the present/recent past construct their infantry groups and how this is likely to change. The United States treats snipers, classically, as lone wolves. We neither train or use many of them - in part for this reason they are regarded as a military romantic image in some circles. They have come to represent the American tradition of pseudo-narcissistic individualism. Alternatively, the Soviets and various Soviet-bloc countries often trained sharpshooters and snipers to distribute among normal infantry groups, embedded in the culture with a cooperative role.
At the present, the United States military is somewhat personnel starved. As a result, while there are supposed to be two fire teams of five men under a Staff Sgt. in an Army squad, our military culture has adapted to have two fire teams of four. If our military deploys to more zones without a general draft we can foresee a further thinning of infantry presence and a requirement for smaller more mobile groups. We might even move to the Canadian model, with two men per fireteam, two fireteams to an assault group and two assault groups to a section. The communist/ex-communist nations of Asia usually run with three man fireteams.
In forty years, there are two simplified possibilities for powers engaged in war across the globe. One, a high population military due to increased general population and decreased opportunity in the private sector - this would make for larger groups of infantry more tightly controlled by the hegemony of command and a reduced value assigned to the individual soldier. The other is a low population military due to artificial decreases in general population (Epidemics, Terrorism, Infrastructure Disaster, Famine) - this would make for smaller groups of better trained infantry more loosely controlled but considerably more valuable. Note that these are very simplified models of the possibilities, they’re not predictions and there is a lot you can flex around. I just provide them because I think they might be useful.