Even if quantum computers can scale and maintain coherence long enough to finish a calculation, their only real use is running Shor's algorithm to break public-key encryption. And there's only one customer for that - the government. So it'll be export-controlled.
Grover's algorithm is nice, but it only gives a quadratic speedup. If you use travelling salesman as an example, for small problems conventional computers are faster (higher clock speed, massive parallelism). And for large problems, quantum computers still take too long. There's a tiny goldilocks zone, that's probably not commercially viable.