Regarding the hardware requirements for validator:
It’s important to understand that Orbs is a very high scale network… this means that unlike a Bitcoin/Ethereum node that run on one machine, an Orbs node is actually a cluster of many machines. This means that a validator would normally create a cluster of multiple machines in some datacenter to run their node (one node = multiple machines).
The easiest approach is of course for the validator to use one of the standard cloud providers for this task and run their node over AWS, Azure, Google Cloud, etc.
Using this approach, when new virtual chains are created by apps and more compute resources are required to run them, the system can automatically spawn more machines in the cloud to support the node and make it more powerful. Orbs nodes actually do this automatically when deployed with Nebula: https://github.com/orbs-network/nebula . The system uses Terraform to allocate the resources automatically on the cloud account of the validator.