Without knowing much about your network topology, it’s hard to give you a good answer. The answer really depends on your network topology and TEM relay configuration. It would help if you could answer these questions:
How many locations are those 16k workstations spread across?
What kind of network connectivity do you have between the sites?
How many relays do you have, and where are they located in your network topology?
Do you have relays at each site?
If you have a fairly robust relay infrastructure, with relays at each site and less than 1000 clients per relay, and good connectivity between relays and clients, you should be able to deploy the 460 MB task staggered over a day (or maybe even less). If you have many clients connected to relays across a WAN, that number will probably be larger. Whether or not you’re using dynamic bandwidth throttling will also be a factor.
You might also want to deploy a smaller subset of the 16k machines (200? 500?) spread across multiple sites first to get a better “feel” for what your infrastructure can handle.
Also, it is my understanding that the stagger feature works by assigning each targeted client a random time between 0 and x to wait before running the task (where x is the stagger time specified).