Microsoft Azure

How Microsoft Azure is structured, which services and AI tools matter most, how its pricing and discounts work, and how to get started through official channels.

Best suited to
Microsoft-centric organizations, hybrid setups and enterprise AI projects
Key services
  • Virtual Machines
  • AKS
  • Azure SQL
  • Blob Storage
  • Azure OpenAI
  • Azure Functions
Microsoft Azure overview banner with a grid of connected network nodes with highlighted squares

Microsoft Azure is Microsoft’s public cloud, and this Microsoft Azure overview explains how it is structured, what it charges for and who tends to benefit most. Azure is closely tied to Microsoft Entra ID, Windows Server, SQL Server, Microsoft 365 and GitHub, which makes it a natural fit for organizations already built around Microsoft technology.

How Azure is organized #

Before looking at services, it helps to understand the containers everything sits in. A tenant is your organization’s Microsoft Entra ID directory. Inside it, subscriptions are the billing and access boundaries, and each subscription holds resource groups that bundle related resources so you can deploy, secure and delete them together.

Getting this hierarchy right early makes access control and cost reporting much easier later. Our guide to Azure tenants, subscriptions and resource groups shows practical layouts for small teams and larger organizations.

Core services #

Service Purpose
Azure Virtual Machines Windows and Linux VMs in many sizes, including GPU-enabled series
Azure App Service and Azure Functions Managed web app hosting and event-driven serverless code
Azure Kubernetes Service (AKS) Managed Kubernetes clusters
Azure SQL, Cosmos DB, Azure Database for PostgreSQL and MySQL Managed relational and multi-model databases
Azure Blob Storage and Azure Files Object storage and managed file shares
Virtual Network, Load Balancer, Application Gateway Private networking, layer 4 load balancing and a layer 7 gateway with optional web application firewall
Azure DevOps and GitHub Actions Repositories, pipelines, boards and CI/CD automation

AI services

Azure’s AI offering has two main layers. Azure OpenAI provides API access to OpenAI models hosted in Azure, with Azure networking, identity and data controls, and it is managed through Microsoft Foundry (formerly Azure AI Foundry), which also offers models from other providers. Azure Machine Learning covers the full ML lifecycle for teams that train and deploy their own models. Prebuilt services for speech, vision, language and document processing sit alongside these.

Microsoft renames and reorganizes these products fairly often, so treat any name as a pointer to the current docs. For a developer-focused tour, see our overview of Azure AI services and how developers use them.

How pricing works #

Azure bills pay-as-you-go by default, with most compute metered per second or per hour and storage metered per gigabyte per month. Common ways to reduce cost:

  • Azure Reservations: one-year or three-year commitments on specific resources such as VMs or database capacity.
  • Azure savings plan for compute: a commitment to an hourly spend that applies across eligible compute services.
  • Spot VMs: discounted spare capacity that can be evicted.
  • Azure Hybrid Benefit: lets you apply eligible existing Windows Server and SQL Server licenses to Azure resources.
  • Free account: new customers can typically get a credit for the first 30 days, some services free for 12 months and a set of always-free services. Confirm the current terms on the official free account page.

Microsoft Cost Management, included with Azure, provides budgets, alerts and cost analysis by subscription, resource group or tag.

Where it fits well #

  • Organizations that already use Entra ID for identity and want single sign-on and conditional access across cloud resources.
  • Windows Server, .NET and SQL Server workloads, especially when existing licenses can be reused.
  • Hybrid setups that connect on-premises environments using Azure Arc, ExpressRoute or VPN gateways.
  • Teams building generative AI features that need enterprise identity, private networking and regional data controls.

Things to consider #

The portal is large, and the same task can often be done in several ways (portal, Azure CLI, PowerShell, Bicep, Terraform). Pick one infrastructure-as-code approach early to keep environments consistent.

Pricing pages can be hard to compare because VM series, regions and licensing all change the final cost. Use the official pricing calculator and review Cost Management after the first few weeks.

Subscriptions have default quotas, for example on vCPUs per region and per VM family. Access to some AI models also depends on region and may need approval. Request quota increases through the Quotas page or a support request in the portal.

Getting started the official way #

  1. Sign up on the official Microsoft Azure website using a Microsoft account or work account. Microsoft verifies your identity with a phone number and a payment card during sign-up.
  2. Enable MFA on the account and, for organizations, set up Entra ID users and groups instead of sharing a single login.
  3. Create separate subscriptions or resource groups for production and testing.
  4. Set a budget with alerts in Cost Management.
  5. Work through the tutorials and learning paths in the Azure documentation on Microsoft Learn.

NewCloudStore is independent and not affiliated with Microsoft; trademarks belong to their owners.

Guides on Azure

All guides

Have a question about a platform or a guide?

Send a note and we will point you to the right overview, guide or official documentation.

Contact us