That's not really true. The servers themselves and their provider may be under US jurisdiction, but not FastMail itself, which means they can take measures against unauthorized access (e.g. encryption) and they can detect breaches and take steps.
Nothing will be foolproof, of course, since the NSA is the worst enemy you could have and if they have physical access to those servers, you simply can't claim you're 100% protected, but it's much better than nothing.
See: https://www.fastmail.com/about/privacy.html
Furthermore, there are actual trade agreements between countries and if the NSA would access the servers of foreign companies on simple whims, the US can lose a lot of money if customers find out. I know of at least one big German multinational company that banned the usage of several US-based services and tightened security, after the Snowden revelations, for fear of industrial espionage.
In other words, the price for breaching users' privacy is significantly higher, because in Google's case, they just have to ask for it and Google can't even disclose such breaches to their customers even if they wanted.
That being said, decentralising a bit to providers that don't specialise in user profiling but more in email handling, is a step in the right direction.
We get what we pay for. Free email is not free. The price is the security that we give up. Gmail and Fastmail are equivalent to me. Both own your privacy that you consciously give up for free...