Understanding Dysprosium Criticality, page-244

BEV magnets can and have been routinely made from impure inputs.
For example, magnetmakers can use NdPr as the base material, where the blend is simply that which they get from a large (and largely homogeneous) mine and extraction facility, which means one does not have to spend the money required to separate. They can also substitute the NdPr with up to 20% of a similar natural occurrence post extraction LaCe product to save even more money. And finally they can provide effective heat resistance for the magnet by doping/treating/GBDing it with a similar product DyTb. Sounds good.
For example the erstwhile 2018 "Toyota magnet" was made using LaCe in a defined ratio that was remarkably close to that which Lynas produced at the time.
But no deposit is so homogenous that you can proceed willy-nilly without concern that magnets made will remain up to the job.

Magnet grades and production technologies have proliferated because it has become increasingly recognized that making magnets to meet more precise performance specifications results in significant resource conservation and reliable performance. You use only what you need and pay for same. So the task of magnetmakers is making magnets that meet precise specs using the cheapest recipe! You don't need a better magnet, you need a cheaper well-defined and sufficient magnet; the recipe can be quite variable.

To find the cheapest blend of REE inputs and technologies that do the job for less not better, highly-defined magmetal inputs are required and to do that you invariably will need at the very least some pure La,Ce,Nd,Pr,Dy,Tb to adjust any blended inputs. Better yet, just have the whole suite of purified REEs available.

Many here have been around long enough to have missed important science and tech developments that have altered the relative values of individual REEs. Specialty products at first, standard practices follow and pure REEs are preferred. You can watch the space by simply googling "new uses for Xx" with time filter; it's quite busy. Terbium is on the rise, for example.

I am happy for Lynas that they will now produce at industrial scale each of the previously identified "magmetals".
This will recreate an ex-China magnetmaker market that can be depended on. It's a whole new universe once they have secure inputs. And those magnets are needed by additional ex-China motor manufacturers that have been concerned about supply chain issues and in some cases abandoned REE magnets.

I suspect Lynas will see a surge in ex-China demand in the near future and that many plans have already been made by Japan and partners that will take advantage of Lynas's unmatched and comprehensive prowess and products.