According to RightScale’s 2018 State of the Cloud report, 66% of enterprises expect to spend at least 20% more on cloud this year. Ironically, those same expectations are almost certainly wildly off, as a whopping 97% admit to not being able to manage cloud costs. In other words, cloud is an even bigger deal than CIOs believe, because their developers keep pushing workloads into cloud providers.

Sending money to the clouds

It’s hard to turn off the flow of funds to the cloud–it’s simply too convenient. Polling close to 1,000 people, RightScale’s survey revealed that more than 25% of enterprises are now spending $6 million or more each year, with plans to grow that spending considerably in 2018:

This sounds great (if you’re a cloud vendor), but the numbers are almost certainly understated, and by a lot. This isn’t RightScale’s fault: How do you measure ad hoc, bottom-up, widespread adoption? We already know not to ask the CIO, who tends to be the last person to know about what’s going on in the ranks, and fortunately only 13% of RightScale’s respondents fit the “executive” description.

It gets no better when asking the real force behind cloud adoption: Developers. Thirty-one percent of RightScale’s survey respondents are developers, with another 18% coming from cloud architects and 41% from IT/Ops (10% come from the business).

Those developers absolutely know what they are spending, but how would they possibly know how much the overall enterprise is spending? Answer: They don’t. Not even close. They may touch up to $6 million in cloud spend, but it’s just as likely that there are significant other pockets of adoption that could tally up to that much or more.

SEE: Cloud migration decision tool (Tech Pro Research)

But there’s another problem with the data: The words “expect” and “plan.”

If we’ve learned anything about the public cloud to date, it’s that developers are largely driving the adoption, and much of that adoption isn’t seen by the CIO until it’s too late for her to stop it. Back in the “good ol’ days,” some technical decision-maker had to approve IT spending. Today, a developer can spin up a new instance on a whim without waiting the six months (or even six minutes) to get provisioned. She may well be planning to roll out a new application on AWS, Azure, or Google Cloud, but cross-enterprise planning? It happens, yes, but the individual developer answering the survey may not be best positioned to answer the question (and, really, no one is in a great position to answer it).

In such a world, convenience trumps most everything else. Including, perhaps, common sense. Or common “cents,” as the case may be.

Overindulging on cloud

Directions on Microsoft VP Wes Miller told me that, at least for Microsoft Azure, enterprise spending on cloud often exceeds the enterprise’s actual ability to consume it.

“[T]he purchasing that matters is at the enterprise level, not ad hoc with individual devs/teams,” Miller said. “There, we see far more orgs that have dramatically over-booked for what they can successfully burn over three years. I anticipate that getting worse.”

Miller is almost certainly correct that significant spend happens at the enterprise level, and that such purchases struggle to figure out how to optimize spending. Small wonder, then, that the top priority for RightScale’s survey respondents (58% of respondents) is optimizing cloud spend in 2018.

That desire for controlling costs, however, is going to keep butting heads with the reality of bottom-up, developer-led adoption.

SEE: Amazon Web Services: The smart person’s guide (TechRepublic)

Fortunately, we’re seeing increasing collaboration between developers/lines of business and overarching IT functions. This collaboration is being driven by fundamental shifts in how enterprises think about technology. For those companies most attuned to cloud, however, the “collaboration” may actually reflect IT losing gracefully. As Rishidot analyst Krish Subramanian argued, “One thing that is common to [companies that go ‘all in’ on cloud] is that they have a relatively small IT team compared to traditional enterprises. This points to the factors slowing down cloud adoption. It is people.”

Whatever the size of the IT department, all companies are having to fundamentally rethink their applications, with cloud-first increasingly a matter of survival. One example I am familiar with is that of a large enterprise that was trying to figure out how to rearchitect a massive application first conceived in the early 2000s. At the time it was first built, the enterprise had very different needs from today–thousands of users, gigabytes or terabytes of data, customers all sitting in the same region, performance important but not all-consuming. This enterprise built on internal servers and focused on a scale-up model. That’s all there was.

Today, that same application has millions of users, distributed globally. The data volume is in the petabytes (and approaches exabytes). Performance latency must be measured in milliseconds and, in some cases, microseconds. There is no option but cloud.

More applications look like this today than the earlier instantiation of that application. Cloud is the only answer, and so IT and developers are working together to realize the vision. Such applications are big enough that they’ll be on the CIO’s radar and will find their way into survey results from RightScale and others.

But even for critical applications, a lot of data on what’s being adopted, and by whom, is lost in the mad rush to the cloud. It’s not the fault of RightScale’s methodology, it’s just a simple fact of how broad the sprint to cloud has become, and how atomized it remains.

So, if you want to know the percentage of companies that are increasing cloud spend, and by significant amounts, in 2018, the answer is “100% of them.” And if you’re wondering how much they’re spending, just look at the income statements for AWS, Microsoft Azure, and Google Cloud. The answer is “low tens of billions today, high tens of billions tomorrow.”