The look angles would be different, but very well-known. So, in practice, we would still be able to use all datasets simultaneously after some pre-processing and geometry corrections.
Scientists combine information from SAR images and Sentinel-2 (optical) to look at and monitor areas of interest. For example, one can obtain information on the location of a newly formed sinkhole (or other hazard) through the optical channel (something you can't see directly from SAR in all cases) and then analyse the deformation in the surrounding environment through the coherence of the inSAR product.
To obtain the inSAR images, one has to apply orbit corrections (which are very important to the correctness of the analysis). I imagine that having the 2 instruments in one satellite would mean that you can do this process once to generate your final product.
So, in general you may use the optical information for land classification and combine it with deformation rates. Optical sensors also don't produce information during the night, so depending on the application, radar can not only be complementary, but the main source of information.
Something else to consider: Maybe this start-up is planning to generate the corrected products already, before sending the info back on Earth, or pre-process it with their models and release them to their clients ready-for-use.
It could also be more cost-efficient to design a single satellite with multiple payloads instead of multiple satellites with one instrument each.