If ITER proves its style to be a success (what may happen this decade, but I'm not holding my breath), we can expect a decade (about half the time of the original) to just complete the first generation of commercially viable reactors (what ITER will not be).

After that, lithium supplied T grows exponentially with the reactor usage, with a doubling time of a bit less than a decade. So, if it becomes the bottleneck (what means, if people to everything else right) you can expect some 3 or 4 decades before fusion even becomes a major power source.