It’s getting harder and harder to remain skeptical when it comes to the value and profit potential of the cloud, and even more difficult to deny that public cloud adoption is maturing beyond just test and development like some cynics had speculated.

Further validating the public cloud’s credibility is new research by Eileen Smith, program manager at IDC, who predicts that the U.S. public IT cloud services revenue will grow from $18.5 billion in 2011 to $43.2 billion by 2016. The market will experience a compound annual growth rate of 18.5% over the forecast period – making it increasingly challenging to argue that the cloud doesn’t have staying power.

The “services and distribution” vertical, which includes retail and professional services, spent the most money on cloud services in 2011, accounting for 30.3% of total revenue that year. Smith also measured cloud spending in financial services; manufacturing and resources; infrastructure services (includes media and utilities); public sector; and home businesses and consumers.

The forecast tracked five cloud service segments, which IDC refers to as: Applications as a Service, System infrastructure software as a Service, Platform as a Service (PaaS), Server as a Service, and “basic” Storage as a Service. It’s interesting to note that most of the services studied by IDC also ranked as high-growth areas in TechTarget’s recent Cloud Pulse survey. Our survey asked 1,497 IT professionals about their use of and attitudes toward public, private and hybrid cloud, and the results indicated that respondents continue to be drawn to public cloud due to several advantages it has over on-premises deployments (such as increased infrastructure availability, scalability, cost management and the “as a Service” model).

Also worth noting is that two months ago, another IDC analyst released a forecast for global public cloud services spending over the same time period. Worldwide, revenue is expected to shoot up to $100 billion by 2016 — meaning that U.S. cloud revenue may account for almost half of worldwide spending. So, if there are skeptics still out there, soon enough you may have to leave the country to hear their arguments.

Wired got an interesting scoop from Intel this week: Google has been buying chips directly from Intel for five years, bypassing traditional data center infrastructure vendors to build its own servers, storage and networking devices.

According to Wired‘s Robert McMillan:

“It’s certainly not anything that the companies involved discussed openly, so it is kind of a big deal that they said something on the record,” says Dean McCarron, who runs Mercury Research. McCarron reckons that about 10 percent of the approximately 16 million server chips sold each year go to big internet companies that are skipping the mainstream server vendors.

One of the reasons that data center operators are so hush-hush about their purchasing decisions is that they don’t want to tip their hand whenever they’re adding capacity. That can drive up prices for other components of the build out — bumping up costs for everything from real estate to networking gear.

Obviously, very few cloud providers have the means or motivation to follow Google’s footsteps here. And while Google’s strategy is grounded in using custom hardware to build a competitive advantage, it echoes something I’ve heard anecdotally from cloud providers big and small: Vendors often aren’t meeting their needs, so they’ve taken to building their own systems.

This is no shocker for the larger providers, particularly the telcos, which have long developed custom systems and applications through their internal research and development divisions. Nor have I heard about a consistent area of complaint among cloud providers adopting a do-it-yourself strategy — sometimes it’s with networking, sometimes storage, sometimes platforms. But it seems those anecdotes and stories like this one from Wired — in addition to the growing interest in open source projects like OpenStack and OpenFlow – all seem to point toward the suggestion that either vendors aren’t satisfying cloud provider needs, or data center infrastructure has become such a differentiator that it’s better to build rather than buy.

Either way, it’s not good news for vendors.