Big data is a different thing from datawarehousing.
In a big data scenario you have lots of data, that you process with a highly scalable solution.
In a databasehouse you collect data from different sources and transform them in several steps to a datamodel you can create reports from it in a simple way. .
And there is the other option you just have to process lots of records from a simular source (measuring data), where you carefully monitor and tune the processing of that data.
The question does not even make a difference between this. Most of the above options are not a single person job if you want it done in a reasonable timeframe(that is before business needs catch up).
The hardware depends on the size of the data. hardware sellers are happy to sell their high-end servers with a scaleable SAN, for a price you can buy a mid class car, or even a price that is not listed.But depending on your needs a simple dell server with a good backup facility under 1000$ might suffice.
no way to tell
PS, "services" might be a way to convince you that that the solution needs to run in the cloud (to complete the buzzworld cicle) or a way to sell consultancy hours.