It’s the end of the year, and along with the grocery lists for parties and the lengthy wish lists from the kids, all good research analysts are building their lists of predictions for 2017. CompTIA will present a full slate of predictions in our upcoming 2017 IT Industry Outlook (due in January), but one in particular sticks out for me as a bridge between trends of years past and trends of years to come.

Typically, a list of predictions will feature major themes that have come up over and over throughout the previous year. Security is a safe bet every year. IoT has been a popular pick recently. Big Data made the headlines a couple years ago before suffering from some backlash. But this year, there seemed to be many topics hovering as potential candidates. Between blockchain, hyperconverged infrastructure, artificial intelligence, bots, and VR, there was an awful lot to talk about.

At the same time, one of the constant themes for most of the past decade didn’t appear to have a specific aspect that would crystallize over the next 12 months. As important as cloud computing is to the shifts happening in IT, the vast majority of companies had bought into the concept and started moving systems to the cloud. CompTIA’s research this year found that companies were digging deeper into this transition, which sometimes led to a more restrained approach as the difficulties of cloud transformation became apparent.

So what was there to say about cloud that hadn’t been said before? And how were all of these new concepts going to fit into the picture? It turns out that the two things go hand in hand. As part of introducing a new era of enterprise technology, cloud creates a new platform. This platform is the foundation for tools such as blockchain or AI. It’s not possible to talk about these tools, or for a company to begin using them, without an established cloud baseline.

Given the speed with which companies jumped into cloud initiatives and the recent research showing a reality check on cloud usage, businesses need to be careful of moving too fast towards new technology. The Big Data backlash was largely caused by companies assuming Big Data was as accessible as cloud, then finding their current data management practices were not solid enough to allow them to jump into bigger things. Without a good foundation in cloud—including both the technical pieces and policy/workflow modification—organizations will not be successful with cutting edge concepts.

Even a good start in cloud may not be enough to immediately learn new tricks. Many of the newer trends take advantage of cloud’s scaling abilities to become far-reaching endeavors. For example, one of the primary IoT use cases is Smart Cities. While there are certainly pieces that can be carved off of these huge initiatives or ways that the technology can be used at a smaller scale, many SMBs will not have obvious applications for these tools until they trickle down in different forms.

From that perspective, the cloud era might look more like the mainframe era than the PC era. Mainframes defined a technical architecture for companies. PCs drove change in the IT function, but in many ways they were an extension of the existing architecture. Cloud and mobile are defining a new architecture. SMBs can easily access cloud to replicate existing functions, but re-architecture (including the use of new tools) is a much larger challenge.

In fact, it might even be accurate to view cloud (probably along with the Internet) as a dividing point in technology history, where each half is defined by a particular platform for creation and a corresponding distribution model. Perhaps the reason it’s getting hard to pick a specific projection for cloud is that cloud isn’t just a part of the IT landscape anymore—from here on out, it’s all cloud.