Also, NSF likes to give large, recurring grants to big institutions, because that maximizes the $/effort ratio. Single-person software projects are below the le

Also, NSF likes to give large, recurring grants to big institutions, because that maximizes the $/effort ratio. Single-person software projects are below the level they like to fund. SBIRs are better for small amounts, but in practice they are so narrowly targeted that unless you lobbied 2 years ago for something to be in there, it won't be.