When you buy through affiliate links in our content, we may earn a commission at no extra cost to you. Learn 
In a wide-ranging interview Sid Nag, vice president analyst at technological research and consulting firm Gartner, says that the cloud model has shifted, saying: “It used to be that your cloud models drove business outcomes — now your business outcomes are driving cloud models.”
There are a lot of new ideas to dig into — starting with the choices you need to make around keeping your data and software on-premises, or cloud-only, or multi-cloud, or hybrid cloud.
After that, how do you find the team to manage your setup? How do you manage costs and keep entire teams from spinning up their own instances?
About Sid Nag
Nag’s areas of focus include public cloud, hybrid cloud, multi-cloud, cloud trends, cloud market insights, cloud product strategy, cloud marketing strategy, go-to-market, cloud migration, cloud networking, and hyperscale edge.
He also provides thought leadership for the Gartner public cloud market share and the public cloud forecast research programs. He is a member of the Gartner Cloud Leadership Council.
Nag is Gartner’s most widely quoted analyst in the press, including Wall Street Journal, Forbes, and Bloomberg. Nag previously held the role of Gartner vendor lead analyst (VLA) for Dell and co-VLA for Google. Nag holds 12 issued patents in the networking and cloud space and is a member of the Gartner Cloud Leadership Council.
Key Takeaways
- Gartner’s Sid Nag: “Cloud was a disruptor — now it’s an innovation facilitator”
- “It used to be that your cloud models drove business outcomes — but now your business outcomes are driving cloud models”
- There will be times when you are running software on Cloud A, another tool on Cloud B, and leaning on Cloud C for a specific best-of-breed feature. How do you deal with data costs, networking, and infrastructure?
- “There are instances where there is a good reason to keep things on the premises, but not always”
- How are edge computing and generative AI shaping the new breed of cloud providers?
Cloud Computing in 2024 and Beyond
Q: What’s coming for cloud computing in 2024 and beyond?
A: Advancements in generative AI technology are going to make a big difference in the way cloud models operate, as you’re going to see a much more sophisticated ability to run applications that are much richer in content.
You’re going to see GenAI being spread across different cloud substrates or infrastructure platforms and software. And I think you’re going to see the cloud being more indispensable and heavily impactful.
Previously, 50% of people thought the cloud was necessary, now, more than 90% of people think that cloud apps will be needed.
I think it’s disrupting business and creating brand-new models. Cloud started out as a mature technology disruptor; then it became more of a capability enabler; then it became more of an innovation facilitator.
And now it’s really a business disruptor. It used to be that your cloud models drove business outcomes, but now your business outcomes are driving cloud models. That’s the shift that has occurred.
Q: What are some of the biggest challenges facing the cloud computing industry today?
A: It all depends on the context, but it’s mostly keeping up with the fast pace of innovation, keeping up with the technology, and keeping up with the skills required in-house to run and manage a cloud deployment model.
Cloud is moving fast and furious, and skilled employees are not available easily. And not all customers, clients, or end-users have the resources for it.
Cloud-First Strategy Is the Future
Q: What is a cloud-first strategy?
A: A cloud-first strategy essentially means that when you’re developing a strategy for your IT services or IT operational model, you should always think of the cloud as an option first before you decide to go in a different direction. That’s the idea of cloud-first.
Q: Why is a cloud-first strategy the future of cloud computing?
A: That’s the future in terms of how you consume IT services. Because if you build out your own data centers, it’s much more expensive. So it’s a shift from a CapEx [capital expenditure] ownership model to an OpEx [operational expenditure] ownership model.
That’s because you’re going to a cloud provider that’s using a common set of resources to run your IT operations or IT environment, as opposed to dedicated hardware, dedicated resources, and dedicated data centers that require dedicated power, dedicated real estate, and things of that nature.
So it’s much more expensive when you build your own data centers versus using a common set of resources where cloud providers have multiple clients running on the same sort of infrastructure.
Corporate Spending on Cloud Computing
Q: Is corporate spending on cloud computing increasing? And if so, what are the trends driving that spending?
A: Corporate spending on cloud computing has always been increasing. Spending is expected to increase to $820 billion in 2025, up from $667 billion in 2024, according to Gartner’s Worldwide public cloud services end-user spending forecast [published in Dec. 2023].
It’s growing exponentially. And, of course, beyond infrastructure, the cloud is a platform for modernizing your applications and workloads, which is very difficult to do on-premises because of the set of tools you need to do that.
For example, if you are developing an application, you need several development resources, whether they’re SDKs, tools, or different programming languages at your disposal.
For you as an IT leader to run and operate your own environment, it can be very difficult, expensive, and time-consuming. So you go to the cloud and get all the resources, whether it’s platform as a service for the middleware type of applications, databases, etc. And again, it’s economies of scale.
Q: How can enterprises simplify their cloud operations and manage/control their cloud costs in the face of escalating IT costs?
A: You have to be intelligent about how you use cloud resources. You want to make sure you have a policy strategy. You want to make sure you have a governance strategy.
For example, you don’t want to give your end users the authority to go and spin up cloud instances because you’re going to get runaway costs with no oversight. You want to make sure that you’re using these cloud resources intelligently.
You want to make sure that – especially if you’re multi-cloud – you have some sort of abstracted layer that allows you to get these compute resources across different cloud providers that have different combinations of infrastructure, compute capabilities, storage capabilities, and networking capabilities bundled into one.
The pricing model could vary based on the cloud provider. Sometimes, the same exact cloud instance could be cheaper in the US versus Europe, given the cloud providers’ regional footprints. So, there are many things you have to worry about when you are looking at the pricing model.
Again, you want to modernize your applications to the fullest degree because when you modernize them, they run more efficiently. And that will also drive lower costs. Those are some of the things you have to do to achieve the benefit of cost savings.
On Hybrid Cloud and Multicloud Environments
Q: Many companies that are migrating to the cloud will likely continue to keep some of their critical applications on-premises. What are the benefits and challenges of operating in hybrid cloud environments?
A: There are instances where there is a good reason to keep things on the premises, but not always. Oftentimes, there are unfounded fears – people are afraid of losing control. They obviously want to protect their own jobs. They have a lack of understanding of the benefits of cloud.
But you may have certain data residency requirements, or you may have certain security requirements that force you to keep certain applications, i.e., your crown jewels, on-premises and not move them to the cloud for various reasons.
So you could move the application to the cloud, but you could keep the data on-premises to satisfy those requirements. That’s an example of a hybrid deployment model.
And there are also instances where you run certain applications on-premises and some on the public cloud. But they always need to be interacting with each other because nothing is an island. So those are reasons why you want to deploy a hybrid model.
The challenges are in how to stitch these things together, which is often difficult because the skills available in most IT shops are not that sophisticated. It’s very hard to find the right personnel who have these skills.
But even if you find the people with the right skills, the rate structure for these technicians is very high, so you may not be able to afford them.
So there may be different reasons why it’s difficult to deploy a hybrid model, which is why 65% of organizations today are essentially employing a third party, such as a global systems integrator provider, to do the job for them.
So, there are the benefits of automation. However, there are challenges for a number of reasons, such as skills, wages, sophistication, and deploying these hybrid models.
Q: How do organizations implement a multi-cloud strategy?
A: The reason you need multi-cloud is essentially that workloads need best-of-breed capability. So I may have a sophisticated application that I’m running on, let’s say, Amazon, which is my primary cloud provider. But I may decide that a particular application or workload needs an analytics engine, let’s say Power BI, which is a best-of-breed solution only available through Microsoft.
So, in that instance, I’ll still use Amazon as a primary cloud provider, but I will lean into Microsoft for Power BI. As an outcome of that, I need to buy some infrastructure services from Microsoft Azure because Power BI runs in the most efficient fashion on Microsoft infrastructure.
And similarly, I could leverage a GenAI engine from Google for that particular application. So now you’ve got an application that’s primarily running on Amazon but leaning on other capabilities from different cloud providers from a best-of-breed perspective. Now, you have a primary, secondary, and tertiary – that is the most prevalent use case for multicloud.
So, how do you actually deploy it? You have to make sure that these applications, i.e., the primary application running on Amazon and the other capabilities from different cloud providers, are talking to each other. That means you need the necessary networking and connectivity model. You need a federated model for data to federate across clouds.
However, there are challenges with data egress because many cloud providers charge for data egress per terabyte of data moved around or the amount of networking bandwidth that’s used to do that.
GenAI/Edge Computing and the Cloud
Q: Does generative AI (GenAI) have the potential to redefine the future of cloud-based solutions and operations?
A: Yes. GenAI is adding to the ability to modernize applications and maintain them. And it has [the capability] to make applications more intelligent, easier, faster, and more efficient to write from an application development perspective.
Q: How does edge computing complement cloud-native architectures?
A: You’re basically taking cloud capabilities and pushing them out closer to your user base. So, the users that are far away from the cloud regional data centers don’t have to reach the cloud regional data centers through some sort of connection. Sometimes, those things can be challenging because not everything is connected. Ideally, they should be, but they’re not.
It depends on the kind of feature parity across the cloud region data center and the hyperscale edge computing platforms. So [with edge computing] you can perform some of the functions that a cloud regional data center would do closer to the user base, [doing away with] such issues as bandwidth latency.
