You're going to have to give some more details on just what the site will do before we can effectively answer you.
You don't need anything as powerful (and as expensive) as Amazon's cloud service until/unless you get at least thousands of clients (the actual point at which you need more than a single server depends partly on how complex your service is).
You can get started with an el cheapo shared server (e.g., GoDaddy.com) for less than $10 a month. But even a dedicated very powerful server will only set you back $200 a month or so.
But there's no doubt that you can get something small and fast and cheap using PHP. I'm just not convinced that it scales well to larger projects. (Though there are plenty of companies out there trying to make a liar out of me.)
And therein we get to perhaps the funniest part of your request:
I ... plan to outsource it's [sic] development to a programmer.
*A* programmer. So if this is any sort of really powerful app, one that presumably lawyers will be willing to pay big bucks for, that means you don't care if it takes 3 to 5 years to get to the testing stage? Because, quite frankly, if you are talking one programmer for a few months, what you will get may be an app, but it's unlikely to be one that will earn big bucks, unless you have some proprietary information you can package with it that will be worth the money.
If you are really serious about this project, you need to be prepared to write a multi-page business plan, which would include detailed specifications of what you expect the app to do [and I do mean detailed! else you will get the trash your would deserve, more than likely], and post it here:
and probably in many other similar places around the web. And then be prepared to be deluged by proposals from India, China, etc., made by one-and-two-person "companies" with no experience in large projects. (Meaning you need to do a thorough job investigating the backgrounds of those who respond.)
Oh... and finally: Were I you, I would *NOT* specify the server-side language/platform. Let the bidders propose it. And then see how many holes you can find in their proposals.