Your point is correct, however the reference to exploiting the...

Your point is correct, however the reference to exploiting the Silex process for UF6 tails is that the process is uniquely advantaged for such tails due to directing its energy (from the laser) into the ²³⁵UF₆ only rather than the whole tails feedstock which (for 0.3% tails) is 99.7% ²³⁸UF₆ - it gets worse for lower assay tails. Because the Silex process largely put its laser energy into the ²³⁵UF₆ component only it's much more efficient than centrifuge (which puts its energy into the whole feedstock) for re-enriching tails. As you say that doesn't stop the process from also have an advantage in enriching to LEU (~4.8% ²³⁵UF₆). Clearly this advantage declines the more ²³⁵UF₆ there is in the feedstock. The other clear advantage of the Silex process is that it is much more flexible than centrifuge in that it can directly produce enrichments assays to client requirements rather than having to be downblended from a higher permanently preset set level and can be stopped (or throttled back) without wrecking the cascade