The question was posed earlier this week as to whether cloud computing ultimately will represent a permanent downshift in IT spending, or whether it will become another channel for growth. Conventional wisdom, as well as Goldman Sachs, suggests the former: While spending will rise in the short-term as companies and cloud providers update their data centers, their investments in virtualization and automation technologies actually will decrease the need for spending in the long-term. It’s a solid theory, but I see the latter theory being equally possible.
The exponential increases in data volumes likely will be a big driver of steady IT spending, although the channels through which vendors make their money might shift. For example, more data and better analytical tools are driving enterprises toward the promise of predictive analytics. While this transition will require short-term software and architectural investments, it also should require periodic expenditures to keep pace with continuous data growth and the desire to analyze even faster.
Should the bulk of data analysis jobs get sent to the cloud, as some predict, software companies still have a potential revenue stream by selling cloud versions of their analytics products. And while profit margins for cloud software and SaaS offerings presently are lower than margins for their in-house counterparts, it stands to reason that increased data and cheaper VMs will mean a greater volume of business; increased demand could warrant higher rates even for pay-per-use services, as well.
The world’s constantly increasing demand for computing power also looks to drive continued IT spending. Granted, much of that spending will be from hosted-service and cloud providers, which means volume discounts on the vendor side and a concentrated pool of options on the user side, but spending is spending. Virtualized or not – and not all hosted servers will be virtualized – service providers will have to buy more servers for the foreseeable future. For the cloud providers, ever-expanding feature sets will mean higher profits associated with each VM they spin up.
Within corporate data centers, new processor designs presumably unavailable in cloud offerings could lead to additional hardware spending, as well as investments in new applications designed to harness these improvements.
Finally, it seems a bit naïve to think that software vendors, especially, will leave well enough alone. Take VMware, for example. Once companies started to realize the cost-savings and (arguably) operational efficiencies of server consolidation, VMware began pushing new tools to drive even further efficiency, improve performance and, ultimately, to build internal clouds that interact with VMware-based public-cloud offerings. Microsoft’s brass is worried about lower profit margins from its move to offering cloud services and, perhaps not coincidentally, it is rolling out the internal AppFabric solution along with the cloud-based Azure. As long as companies maintain their own data centers, there always will be opportunities improve efficiency, performance and flexibility, and software vendors will continue to exploit these openings with new products.
It’s possible that cloud computing will cut the legs out from under IT spending permanently, but it seems unlikely. Assuming vendors never get IT entirely perfect, and assuming companies still have IT budgets to spend, there should be plenty of revenue to go around – and the cloud might even open lucrative new streams from which to derive it.