What’s the compute for?
I can see a small amount of cpu being used for data transformation, but these are pretty decent sized boxes and realistically on-prem compute probably could have met the requirement in 99% of the cases anyway given the lead times for delivering the storage to AWS.
If, on the other hand, it was meant for either wan optimisation or as a combination of sending a point in time snapshot to AWS and then using that to seed a rapid data migration, it might make sense.
While there were reports of something to scare Cisco earlier this year, that’s more likely to scare Riverbed/Silver Peak....