That's a great and important question. (doubly so for me, as constrained resources may see me leave the field, to which I have devoted my entire professional life, in the coming months) We are perennially grateful for the support we receive to do the work that we do.
To a high degree of approximation, Einstein's predictions are very, very correct. We are trying to look very carefully at Nature because we don't think we have the whole story.
It is possible that Einstein is exactly right. If so, we would never be able to describe the four forces of Nature with one unified theory. Today, we need two theories to describe everything we see -- the quantum-mechanical/particle-physics description of the Standard Model and gravity. There is no quantum-mechanical fuzziness in the mathematics of gravity, and there is no hint of the differential geometry of gravity in the mathematics of the Standard Model. The aspiration of almost every fundamental-physics physicist is to find a way to either simplify the Standard Model or connect it with gravity.
If the return on investment seems insufficent, know, too, that the technology we develop to push the boundaries of knowledge has important spin-offs. GPS is impossible without corrections from both special relativity and general relativity. The instrumentation we develop to make gravitational experiments possible on earth requires the development of new classes of seismometer [1] that may open new understanding of Earth's dynamics and allow better chip-fabrication instrumentation. The instrumentation developed to test gravity in space is also being used to measure the movement of mass (i.e. ice and water) on Earth's surface [2].
Even more important is the training we provide to students. Just as athletes train in the gym to get stronger, confrontation with the hardest known technical problems provides an efficient path for students and young faculty to become proficient at the entire range of modern measurement science. Alumni from our group have not only gone on to fancy academic positions, but also helped to redefine the kilogram, designed upgraded digital calipers used by tens of thousands of people worldwide, built key components of a major quantum-computing company's infrastructure, and more.
The people who work on these kinds of experiments are making a fraction of what they could make in industry. We are doing it for love, not money, and the returns to society are myriad.
[1] https://arxiv.org/abs/1707.03084
[2] https://gracefo.jpl.nasa.gov/