Cloud infrastructure spending is growing

I hate to say, “I told you so,” but the significant downturn in cloud spending that many people anticipated has not yet happened. Last year and this year, I’ve had one argument after another with tech reporters who all believed that cloud spending would take a nosedive after years of rapid growth. Their reasons included an expected recession and economic slowdown, which made their predictions easy to frame.

Making the opposite call was easy for me as well, but for different reasons. Consider the rise of cloud usage to support AI-based systems and complex application deployment over the past few years. That stuff needs to run somewhere, and the cloud is the best bet. Therefore, I predicted no downturn in cloud spending anytime soon.

The numbers

According to IDC, spending on compute and storage infrastructure products for cloud deployments went up almost 8% year on year for the second quarter of 2023 to $24.6 billion. It is worth noting that shared cloud infrastructure accounted for a significant portion, surpassing non-cloud infrastructure at $14.4 billion and contributing to almost half (45.8%) of the total infrastructure expenditure.

IDC anticipates that shared cloud infrastructure spending will reach a staggering $72 billion in 2023, reflecting the growing reliance on shared computing resources. This emphasizes the industry’s focus on efficiency, scalability, and cost-effectiveness through shared cloud infrastructure.

I suspect this is the start of a second boom in cloud spending. Artificial intelligence is a massive motivator for cloud spending, and many of the applications we’re deploying, such as cloud-native and serverless, are storage and compute hogs. Although there may be some consternation about this level of spending, most enterprises will view it as unavoidable. They will find the money somewhere to get to the level of cloud scale needed to support the next generation of “mega-cool” applications and systems.

Not so fast

Although I love to be right about the strong cloud spending, that does not mean it’s suitable for all enterprises. Indeed, the trend will be to overspend, even after net-new finops deployments that closely monitor where the dollars are spent.

We must focus on accountability, automation, and discipline around allocating and paying for cloud resources. I suspect many cloud deployments are hugely underoptimized and need a tune-up. Even though some of this shared infrastructure spending is unavoidable, CIOs need to review how the spending occurs and look for opportunities to save dollars without reducing the value generated by these systems.

I suggest companies consider all other options, such as bringing some processing into enterprise data centers. Those prices have been falling while they have been stable or rising on the public cloud side. Also, many systems function in isolation and don’t benefit much from existing within a public cloud. Simple storage is one example, and many enterprises are putting those systems on-premises these days. Companies that go that route are not rejecting cloud-based platforms but rather looking for better, more cost-effective options. Of course, you need to consider the more significant value of leveraging a public cloud, including innovation and connected services.

The future

The future of all this is easy to see. Shared infrastructure spending will continue to rise, as the IDC report predicts. I agree with that. Enterprises are still willing to toss money at problems, in this case, tossing money to the public cloud providers.

However, I hope that enterprises become savvier in terms of managing that spending better moving forward. The tools available to us, such as finops systems and AI, allow a deeper understanding of what we’re spending and for what purpose.

I also hope enterprises review current architectures in terms of efficiencies. We’re leaving millions of dollars on the table, which can be fixed with configuration and platform changes. In other words, we’re paying for our past mistakes when we rushed into the cloud. Time to fix them.