You're right though, "the cloud" is siphoning up a lot of money that would be going to others.
I do wonder how much damage a cloud-everything approach does in start-ups when it goes wrong.
For one thing, I've seen a few early stage stories where businesses seemed to be spending more per month on the fees for the 20+ online services they use than they were spending on hiring good people. They get locked in and dependent, and then their costs are ongoing and scale up with things like head count. A few years ago you might have bought a handful of servers and hired an IT guy to do the same kind of things, and maybe you'd also build-or-buy a small number of software tools for admin purposes.
The usual arguments against that old school approach seem to be that going to the cloud outsources most of the admin work leading to better security/reliability and that it allows better scalability, but I don't really buy either of those as a blanket statement.
In terms of admin, you still have to both identify the cloud services you want to use, in possibly the most hype-driven market I have ever seen, and spend the time to set them up. Those are not small overheads.
In terms of reliability, as we've seen with things like Amazon outages, even relatively well-established businesses who you'd think would know better still don't actually get their resources set up to be properly resilient under pressure, so the idea that you no longer need smart in-house people to run IT just doesn't hold water.
And in terms of scalability, while you certainly can scale almost in real time with some of these services, how many start-ups really need to scale faster than they could rack up some new servers and networking gear of their own with a standard image installed? The downside of paying much higher long-term costs for basic computing and storage resources seems to kick in relatively low with a lot of cloud pricing today.
Obviously there have been some very successful online startups in recent years, both in B2C and B2B spaces, but I do wonder sometimes whether a lot of these XaaS cloud providers are really providing the clear advantages that so many people keep talking about.
In part because each of those overvalued and overfunded startup were buying Sun servers, Cisco switches, EMC storage, etc. Now they're renting some EC2 instances or whatever. As a result, when the bubble deflates at least a bit, the impact is hopefully largely limited to some VC portfolios, premium developer salaries in some geos, and maybe the income flow for some northern California landlords (or those who paid $2M for a fixer-upper in that area). Not especially pleasant for those involved to be sure, but hopefully not a huge deal for everyone else.