The quantum computer company Alice & Bob’s claims they are on track to have by 2030 a fault tolerant quantum computer with 100 logical qubits. I don't know if they will reach their goal but it doesn't seem to be beyond the realm of possibility, and if they do it will change the world. Such a machine could accurately simulate the ground state energy of moderately complex molecules such as iron/sulfur clusters that are important in biological enzymes. It could simulate small superconductors, magnetic materials, and high-temperature catalysts. It could speed up data search thanks to Grover's algorithm, and optimize logistics and scheduling problems.
One specific example is it could simulate nitrogenase, or at least the active electrons in it that are important for bonding, and nitrogenase is the enzyme that plants used to break the enormously strong triple bonds in nitrogen molecules in the air so they can be used by biological organisms. Plants are able to do that using very little energy, but nobody has a good understanding of how they do it, so we have to use the Haber–Bosch process which is so inefficient it uses about 2% of the human race's entire energy budget, but without it hundreds of millions of people would starve to death. It would be nice if we had something better.
And of course such a machine would be very useful for cryptography. But probably the most important things of all are things we currently don't know anything about and will not until we have a large low fault quantum computer that we can fool around with.