What is cloud computing?

The term may sound more abstract than it actually is. At its core, cloud computing means providing and using IT resources over a network, on demand and with measurable usage. These resources include:

• Computing power and virtual machines
• Storage and databases
• Networks
• Software and development platforms
• Analytics services and, in recent years, AI services

Cloud computing is a form—and an evolution—of outsourcing: Instead of operating every resource themselves on a permanent basis, businesses can obtain, expand, or scale back resources as needed. This is a key difference from traditional IT: Cloud computing changes not just where IT operates, but how businesses obtain and manage IT resources.

Not every IT environment running in a data center automatically qualifies as cloud computing—even if it can sometimes seem that way today. Several key terms are commonly used to describe the cloud:

• On demand: Resources can be used as needed—obtained and released at short notice. This highlights their nature as a service, much like getting a haircut or eating at a restaurant. That is why cloud resources are often described as being used “as a service.”
• Automation: Individual services can largely be provisioned automatically at the push of a button—whether they are started or stopped manually through a console or through an API call. Consoles are still common in smaller deployments, but cloud automation generally goes much further: Most cloud workloads today are managed through containers or automation platforms.
• Scalability: Capacity can be increased or reduced flexibly. Ultimately, scalability combines on-demand availability with automation. T Cloud Public offers a service called Auto Scaling: It increases or decreases server capacity based on a service’s workload. Users must define specific instructions, known as policies, for it to follow.
• Standardization: Resources are accessed through standardized interfaces and services. Strictly speaking, standardization is not a must-have requirement for clouds. It is more of a necessity for providers. Standardization creates a consistent technical foundation, making cloud resource pools easier to manage.
• Measured usage: Consumption and resource usage can be tracked transparently. Measurability may be the cloud’s defining characteristic. Measuring resource usage gives users a new level of transparency into their IT consumption. Precise resource measurement provides the basis for usage-based billing, much like electricity or water, while also offering visibility into a service’s carbon footprint.
   

The cloud is more than someone else’s computer. What matters is the operating and provisioning model.

With SaaS, users access a ready-to-use application as a service. The provider handles operations, infrastructure, and usually maintenance and updates as well. SaaS services primarily target IT end users. According to Gartner, SaaS accounts for the largest share of the cloud market, at approximately 40 percent.

The further you move from IaaS toward SaaS, the more responsibility the provider assumes for technical operations.

Cloud computing can offer businesses economic benefits, particularly where IT requirements change rapidly and frequently—or where they need to try things out quickly. It also enables businesses to use IT resources without taking on the work of operating their own IT infrastructure—the traditional make-or-buy decision.

Clouds also provide access to IT resources that businesses do not currently have in-house or that would take too long to procure. This is particularly evident at present, in late 2026, with GPUs, which are in short supply and high demand in the European market. Making IT resources available “like electricity from an outlet” creates business agility and reduces investment risks.

Although more and more applications now run permanently in the cloud, its traditional strengths lie primarily in temporary and fluctuating workloads, such as high-performance computing projects, software development, analytics, and testing.

Defining cloud computing is also a good opportunity to dispel a persistent myth: Cloud computing does not automatically mean lower costs. Its economic value depends on how resources are used, managed, and operated.

In the early days, cloud providers liked to promote the blanket claim that “the cloud is cheaper.” Over time, as businesses gained their first experience with cloud deployments, a more realistic view emerged: “The cloud is not always cheaper—sometimes it is even more expensive.” Additional costs arise not only from “cloud waste”—resources that are provisioned but not used—but also because the OpEx model is not ideal for continuous operation. After all, a race car is not the right vehicle for transporting crowds of students to school every day.

“Flexibility comes at a price” became the next stage in the discussion of cloud costs. In other words, the cloud’s business benefits can justify higher costs than a traditional on-premises deployment. In 2026, the conclusion is clear: Businesses should choose the right cloud wherever it delivers the greatest business value. Cloud costs should be managed through effective cloud FinOps—and the cloud really can be cheaper 😉.

An important aspect of cloud use is that measurable service consumption translates into dynamic pricing models for users: Billing must reflect actual usage—“pay as you use” or “pay as you go.” T Cloud Public calls this “elastic.” Many cloud providers also offer discounts for long-term resource usage, such as T Cloud Public’s reserved models.

No. Traditional IT infrastructure also frequently runs in a data center. What matters is not just the physical location, but how IT resources are provided and used. For example, a business can operate its own servers in a data center without using cloud computing. Conversely, internal IT provisioning can offer typical cloud characteristics such as self-service, automation, and dynamic scaling.

This also explains why asking “Where is the cloud located?” does not go far enough. Questions that are at least as important for businesses include: Who operates the infrastructure? How are resources provisioned? How flexibly can the environment be changed? What management capabilities does the cloud offer? And what dependencies does it create?

The cloud is also frequently compared with hosting, particularly in terms of pricing. The boundaries between these two provisioning models are fluid. The key difference is that a cloud provides users with a set of management services that make IT resources easier to use and manage, with greater automation.

Cloud computing does not automatically make IT requirements easier to meet. Its new flexibility also creates new responsibilities. Businesses moving to the cloud usually discover this during migration. The initial assumption that it is simply a matter of “moving to another virtual server” falls short.

Adopting the cloud requires a shift in thinking. This starts with identifying suitable cloud instance types and architectures that deliver performance comparable to the physical hardware previously used, and extends to establishing a cloud-native culture. This includes:

• Managing costs and resource consumption transparently
• Addressing data privacy and compliance requirements
• Defining security and access models
• Assessing dependencies on individual providers and technologies
• Placing data and applications in the cloud environments best suited to them

Particularly for business-critical applications, the starting point should therefore not be “Cloud—yes or no?” A more useful question is: What are the application’s requirements, and which operating model meets them best?