Spatial join performance and scalability is sensitive to tessellation geometry, particularly when dealing with polygons and other non-point geometries or if you have high precision requirements. The abstract join algorithms are similar but it is much more difficult to make them rigorously correct in the general case with hexagonal tessellations, and they'll never be particularly fast at scale.

One of the main reasons that equal-volume cubic tessellations have emerged as the default choice for high-scale analytical DGGS is that they are nearly optimal for scaling out spatial joins between arbitrary geospatial data models. And relatively optimal in most other regards as well, especially computationally; the primary "downside" is that they are 3-dimensional, which is slightly wasteful, though more and more geospatial analysis applications make good use of a direct 3-dimensional representation.