My solution to this type of problem is to have the government provide a one-time grant for such an institution to be constructed/implemented, and then have no further intervention from the government.

This could be done as a prize paid out to the first one to accomplish some important milestone, such as the privately-funded "X Prize" which paid $10 million to the first team that could send a three-passenger vehicle into space twice within a two-week time period. That is perhaps the best way to make sure that the 'subsidy' is truly a one-time grant (and also to ensure that the government is picking who will win, but simply letting open competition pick the winner).