Cloud deployment models: public, private, hybrid, community and multi-cloud compared

A cloud deployment model describes who owns the cloud infrastructure, who can use it and where it runs. NIST SP 800-145 defines four models: public, private, community and hybrid cloud. Multi-cloud, meaning two or more public providers, is usually added as a fifth. Most enterprises run a hybrid mix and decide per workload.

Key takeaways

• Public cloud is shared, provider-run and pay-per-use; private cloud is dedicated to one organisation; hybrid combines them with shared management and data movement.
• Community cloud serves a group with shared requirements, such as government bodies. Multi-cloud means several public providers, which is not the same as hybrid.
• Hybrid is the most common enterprise setup: 73% of organisations ran hybrid estates in Flexera's 2026 State of the Cloud survey.
• Deployment models (where it runs) and service models such as IaaS, PaaS and SaaS (what you get) are separate choices.
• Choose per workload, based on data sensitivity, demand pattern, integration needs and cost, and review the choice at least once a year.

What is a cloud deployment model?

A deployment model answers three questions: who owns the infrastructure, who is allowed to use it, and where it physically runs. It is a different question from the service model. IaaS, PaaS and SaaS describe how much of the stack the provider manages; the deployment model describes where and for whom it runs. You can run IaaS in a private cloud or buy SaaS that sits on a public cloud. Our guide to the four types of cloud services covers service models.

The reference definitions come from NIST Special Publication 800-145, The NIST Definition of Cloud Computing. They are short and still widely used:

• Private cloud: "provisioned for exclusive use by a single organization comprising multiple consumers (e.g., business units)". It may be run by the organisation or a third party, on or off premises.
• Community cloud: "provisioned for exclusive use by a specific community of consumers from organizations that have shared concerns".
• Public cloud: "provisioned for open use by the general public" and it "exists on the premises of the cloud provider".
• Hybrid cloud: "a composition of two or more distinct cloud infrastructures (private, community, or public) that remain unique entities, but are bound together by standardized or proprietary technology that enables data and application portability".

Cloud deployment models compared

The two tables below work as a diagram of the models: the first shows who owns, uses and runs each one, and the second shows the trade-offs.

Public cloud

In a public cloud, the provider owns the hardware and shares it between many customers, each isolated from the others. You provision resources yourself and pay for what you use. AWS, Microsoft Azure and Google Cloud are the largest public clouds, and all three run regions in Sweden and India.

Choose it for: workloads with variable demand, new cloud-native applications, development and test environments, and teams that want to avoid buying hardware.

Be careful with: personal data that may be accessed from outside the EU or India (this needs a legal review, not just a region choice), and steady, always-on workloads, where you should compare commitment-discounted cloud prices with owned hardware.

Private cloud

A private cloud is dedicated to one organisation. It can run in your own data centre or be hosted by a partner. What makes it a cloud rather than a server room is self-service, pooling and automation, usually through VMware, OpenStack, Nutanix or Kubernetes platforms such as Red Hat OpenShift.

Choose it for: regulatory or contractual requirements for physical isolation, legacy systems that cannot move without a rewrite, and steady, high-utilisation compute where owned hardware stays busy.

Be careful with: the full cost. Power, cooling, hardware refreshes, licences and the staff who patch firmware all belong in the comparison. Private estates also drift from security baselines unless someone actively maintains them.

Hybrid cloud

Hybrid cloud connects private infrastructure with one or more public clouds so workloads, data and management work across both. The connection is what matters: two separate estates with nothing in common are not a hybrid cloud in the NIST sense.

It is the most common enterprise setup. In Flexera's 2026 State of the Cloud survey (published March 2026), 73% of organisations reported running hybrid estates.

What are the benefits of a hybrid cloud deployment model?

• Gradual migration: move systems one at a time instead of in one risky cut-over.
• Placement by sensitivity: keep regulated data on private infrastructure while customer-facing services scale in the public cloud.
• Burst capacity: run the baseline on-premises and add public cloud capacity for seasonal peaks.
• Disaster recovery: use a public cloud region as the recovery site for on-premises systems.

The cost is complexity. Private links such as AWS Direct Connect or Azure ExpressRoute become critical paths that need monitoring. Firewall rules, security groups and network security groups describe the same intent in different languages, so manage them as code. Without one monitoring view, incidents get chased across several consoles. For more, see what a hybrid cloud is and our private and hybrid cloud services.

Multi-cloud

Multi-cloud means using two or more public cloud providers. It is not the same as hybrid. AWS plus Azure is multi-cloud; AWS plus an on-premises VMware cluster is hybrid; doing both is hybrid multi-cloud.

Much multi-cloud is accidental: one team chose AWS, the data team chose Google Cloud for BigQuery, and an acquisition brought Azure. Deliberate reasons include best-of-breed services, regional coverage and reducing dependence on one supplier. Each extra provider adds a different identity model, billing format and skill set, so add one only when you can name the benefit. Read more in is multi-cloud worth it?

Community cloud

A community cloud is shared by organisations with common requirements, such as security rules or a public mission. Government clouds are the clearest example, including AWS GovCloud (US) and Azure Government in the United States. In Europe, the model also appears in shared platforms for public sector bodies and research networks.

Azure deployment models: public, private and hybrid on Microsoft

Many people search for this in Azure terms, so here is how Microsoft's products map to the models.

• Azure public cloud: the Azure regions themselves, such as Sweden Central or Central India, shared by all Azure customers.
• Private cloud on Azure technology: Azure Local extends Azure capabilities to hardware in your own locations, including sites that need data kept locally for sovereignty or regulatory reasons.
• Hybrid: Azure Arc brings servers, Kubernetes clusters and databases outside Azure, on-premises or in other clouds, under Azure management. Azure Local uses Arc as its control plane.
• Community: Azure Government for US public sector customers.

So is Azure private cloud or public cloud? Azure is a public cloud, but Microsoft sells the tools to extend it into private and hybrid setups. The same is true of AWS, with AWS Outposts; see is AWS a private or public cloud?

How to choose a cloud deployment model

Make the decision per workload, then let the overall picture emerge.

1. Classify each workload. What data does it hold (personal, financial, health)? Is demand steady or spiky? Does it depend on on-premises systems? Does it need special hardware such as GPUs?
2. Map the constraints. List applicable rules (GDPR, NIS2, India's DPDP Act, sector regulators), existing hardware with useful life left, contracts and team skills.
3. Pick a model per workload. If everything lands in public cloud, you are public-only. If some stays private, you are hybrid. If several public providers are involved, you are multi-cloud.
4. Review yearly. Prices, regulations and workloads change. Align the review with contract renewals and audits.

Example: from one data centre to a multi-AZ public cloud

ET Network hosted its digital services from a single data centre, with limited redundancy for a growing customer base. Opsio designed and built a cloud-native architecture on AWS across multiple Availability Zones, with automated provisioning, containerised workloads and security controls, and worked alongside ET Network's team from architecture through deployment and testing. Read the ET Network case study.

How Opsio helps

Opsio, with offices in Karlstad and Bangalore, helps organisations decide which workloads belong in public, private or hybrid environments, then migrates and runs them 24/7 through our managed cloud services.

Frequently asked questions

What are the 4 cloud deployment models?

NIST SP 800-145 defines public, private, community and hybrid cloud. Multi-cloud, the use of two or more public providers, is often treated as a fifth model in practice.

Which cloud deployment model is most common?

Hybrid cloud. In Flexera's 2026 State of the Cloud survey, 73% of organisations ran hybrid estates that combine public and private clouds.

What is the difference between Azure public cloud and Azure private cloud?

Azure public cloud runs in Microsoft's shared regions. A private cloud built on Azure technology, such as Azure Local, runs on hardware in your own locations and serves only your organisation, managed through Azure Arc.

Is hybrid cloud the same as multi-cloud?

No. Hybrid combines private and public infrastructure. Multi-cloud combines two or more public providers. An organisation can be both at once.

Is hybrid cloud cheaper than private cloud?

It depends on the workload. Hybrid usually costs less for variable demand because you do not size private hardware for the peak. For steady, high-utilisation workloads, private infrastructure can cost less per unit of compute. Model both for your actual workloads.