Talking about Cloud to host your application, we're comparing to classic hardware. Such servers are large memory—high-CPU computers in a self-hosted datacenter. Many applications, even the high-load ones, running in such a context are monolithic, single-server applications. The Cloud is based on smaller low-cost machines, which are available by dozen when necessary.
Vertical scalability is the option to scale up the application to a larger server when it requires more resources. This makes sense in the Cloud context as the on-demand resources and automated APIs let you restart your application on a new server within minutes. So, you can use a larger server as your application load grows, but keep in mind that there's a limit. Let's consider the features of Amazon EC2 m1.large
computer:
64 bits CPU 4 compute unit (this is equivalent to a 4 Intel Xeon 2007 1 GHz CPU)
7.5 GB memory
800 GB ephemeral disk
Moderate network performance
Such a large server is comparable to the MacBook pro...