I have an issue with this type of question. Given compute moves to commodity, the only valid answer is 'price'. Granted, things like support and uptime need to be factored into cost calculations, but given we're moving to more and more transitional instances this will matter less and less over time.

One advantage I see is if you are away from New York and SF, and want low latency. They have locations in Dallas, Chicago, Atlanta, and Seattle.

Because its a commodity space where you should be using the cheapest provider?