As of 2023, it was apparently <$1000/sequence [0], which means it would be a few hundred million to maybe a billion dollars. Which is a lot, but it really only needs to be done once, and is a global project. That seems pretty reasonable all things considered. But also, we can probably wait a little while since it's not super crucial until artificial genetic modification starts to get widespread, which is still probably a decade off or so. We could probably start just collecting samples now, and as long as they are appropriately preserved, the actual sequencing could be done later when it gets even cheaper.
It occurs to me that the sequencing itself probably wouldn't be the main cost. Collecting the samples, and the manpower needed to do it, would be far, far more expensive. Probably at least 10x the cost, especially if you were really serious about getting the full range of diversity.
[0] https://3billion.io/blog/whole-genome-sequencing-cost-2023