Virtual private clouds once functioned as a secure space inside someone else's cloud, but they have become the connective tissue for hybrid cloud systems that power AI workloads.
Virtual private clouds (VPCs) are no longer just isolated enclaves in a hyperscaler's data center. They have become the networking glue connecting applications, data and AI workloads across public clouds, private infrastructure and on-premises systems. As enterprises race to deploy generative AI and manage sprawling multicloud environments, VPCs are shifting from a configuration exercise to a strategic foundation.
According to IDC, the public cloud infrastructure-as-a-service networking market grew 18.7% in 2025 as enterprises prioritized networking for AI, and the research firm expects the market to double in revenue by 2030. Future Market Insights projects the virtual private cloud market will grow from $55.7 billion in 2025 to $185.7 billion by 2035, a compound annual growth rate of 12.8%. Gartner forecasts public cloud services spending will reach $1.48 trillion by 2029, fueled by AI integration and modernization efforts.
AI is accelerating the need for IT resources as models and agents must access databases, storage, APIs and business applications, often without sending sensitive traffic over the public internet. That demand is putting a brighter spotlight on private connectivity, tighter network controls and the role VPCs play in keeping distributed environments connected and secure.
What Is a Virtual Private Cloud?
A virtual private cloud (VPC) is a secure, logically isolated network that organizations create within public or private cloud environments to run applications, store data and control how resources connect and communicate.
Companies commonly associate VPCs with hyperscalers such as Amazon Web Services (AWS), Microsoft Azure and Google Cloud, but organizations also use them in private and hybrid clouds. VPCs let companies run applications, store data, host websites and support other workloads while controlling how those resources connect and communicate.
In a public cloud, think of a VPC as a private, secured suite in a large building. The landlord manages the property, but the tenant controls who can access the equipment and records inside its space, and when. In private and hybrid environments, the same basic idea applies, but the organization may also own or manage the underlying infrastructure.
VPCs are a core networking component of infrastructure-as-a-service (IaaS) environments. Companies can divide them into subnets for different applications, workloads or security needs, then set IP address ranges, routing rules, gateways and security policies to control how resources communicate and which users, systems and services can reach them.
How Do VPCs Work?
VPCs use software-defined networking (SDN) and security controls to create isolated environments, direct traffic and govern how resources connect with one another and with outside networks and services. Common building blocks include:
IP address ranges and subnets: Companies assign IP address ranges to VPCs, defining the network space available to their resources. They then divide that space into subnets, or smaller network segments. This lets organizations group and separate apps and workloads, control how traffic moves between them and apply different security rules where needed.
Route tables: VPCs use route tables to direct network traffic. Their rules determine whether traffic stays local, moves between subnets or reaches other networks.
Security controls: Cloud service providers (CSPs) usually offer built-in security controls for VPCs. Network access control lists set inbound and outbound rules for entire subnets, while security groups, firewalls and similar tools apply more granular controls to individual resources or groups of resources. Together, they support least-privilege policies by limiting access for users and systems to what's needed for specific jobs and operations.
Gateways and connectivity: VPCs use networking services to link resources with the internet, other VPCs and on-premises systems. Internet gateways provide public connectivity, while network address translation (NAT) gateways translate private IP addresses so internal resources can reach the internet without accepting unsolicited inbound connections. VPNs and dedicated links connect VPCs with other private networks.
Private endpoints and service connectivity: VPCs use private connections to securely reach managed cloud services such as storage, databases, APIs and AI platforms. Private endpoints, or dedicated connection points that employ private IP addresses or network interfaces, give VPCs a controlled path to those services over the cloud provider's private network.
Flow logs and DNS services: VPC flow logs contain records of network traffic, giving IT teams visibility into connections, traffic patterns and potential configuration or security problems. Domain name system (DNS) services translate resource names into IP addresses and support private naming schemes that help internal resources find one another without exposing those records publicly.
What Are the Benefits of VPCs?
Companies that adopt VPCs typically do so because they combine the scale and convenience of the public cloud with the control, manageability and security of a private environment. Benefits include:
Security and isolation: VPCs let companies create logically isolated networks for their applications and data, separating them from other cloud customers and limiting unnecessary exposure to the public internet. They also let organizations segment workloads so a security problem in one area is less likely to spread laterally across the network.
Control and governance: VPCs give IT teams control over addressing, routing, segmentation and public exposure. Administrators can also apply common policies across multiple VPCs, helping to maintain consistent configurations while addressing regulatory compliance requirements.
Private access to cloud and AI services: Like express toll lanes on busy highways, VPCs provide controlled paths to databases, storage, APIs and AI platforms, avoiding the public internet. Keeping that traffic on private networks reduces exposure as sensitive data moves between cloud services.
Hybrid and multicloud flexibility: Technical limitations often keep IT teams from placing applications and data where they make the most sense. VPCs help bridge public cloud, private cloud and on-premises environments, giving organizations more freedom to choose where workloads run while supporting phased migrations, legacy systems and existing infrastructure investments.
Scalability and workload support: As resource demands grow, VPCs let organizations expand their cloud environments without rebuilding the underlying network from scratch. They can add compute, storage, subnets and other resources while maintaining existing routing and security policies. VPCs also support a range of traditional and cloud-native workloads, including virtual machines (VMs), containers, serverless apps and managed services.
Availability and fault tolerance: VPCs support highly available architectures by letting organizations distribute workloads across multiple availability zones. Paired with redundant resources and failover mechanisms, this reduces dependence on any single location and helps applications stay available when individual components fail.
Who Is Using VPCs?
Companies across industries deploy VPCs as part of broader cloud modernization efforts rather than as standalone networking projects. Examples include:
Bedag: Swiss IT service provider Bedag Informatik AG is building a modern workplace environment for employees of the Bern Cantonal Administration. Based on the Nutanix platform, the environment includes a VPC with software-defined networking (SDN), Citrix virtual desktop infrastructure (VDI) and Windows and Linux virtual machine (VM) workloads. (Bedag is a Nutanix customer.)
Cvent: Global meetings and events technology provider Cvent manages hundreds of AWS accounts, with separate VPCs supporting production and development. It uses Amazon VPC Lattice to connect services across those VPCs while limiting network exposure as the environment scales.
Orion Health: New Zealand's Orion Health, a healthcare software provider, built an internal generative AI chatbot that searches more than 500,000 internal records. The system runs inside an Amazon VPC and is used internally. Orion plans to extend the bot to customers.
TRIAL Company: Japanese retailer TRIAL Company is using Google Cloud VPC subnets as part of a modernization effort moving legacy systems from its data center to the cloud. Its hybrid architecture lets on-premises and cloud subnets operate as one logical network, allowing workloads to migrate without changing existing IP addresses.
What Challenges Come With VPCs?
The same controls that make VPCs useful can also make them harder to manage at scale. As deployments spread across accounts, regions and clouds, networking complexity and costs can rise quickly. Key challenges include:
Complexity and VPC sprawl: When different teams design VPCs independently, they can end up with inconsistent addressing, routing, DNS and security policies. As those environments multiply, configurations drift, visibility suffers and troubleshooting gets harder.
IP address management: VPCs need enough IP addresses to accommodate new workloads and subnets, and their address ranges shouldn't overlap with connected networks. Duplicate ranges can create routing conflicts, while limited address space can constrain future expansion.
Cost management: VPCs cost very little, but the services around them can add up over time. NAT gateways, private endpoints, firewalls and data transfer all generate charges that rise with traffic and network complexity.
Security misconfiguration: Logical isolation doesn't guarantee secure configuration. Overly permissive rules, public endpoints, poorly designed routes and weak segmentation can expose resources or disrupt legitimate traffic.
How Should Organizations Get Started With VPCs?
Building a VPC requires upfront decisions about workloads, connectivity, security and scale.
Start with workloads and architecture: Industry experts at AWS, Google Cloud, Microsoft Azure and Nutanix recommend beginning with the applications, users and services the environment needs to support. Determine which workloads need to communicate, which should remain isolated and where they should run. Those decisions shape whether to use one VPC or several and whether they span public cloud, private infrastructure or both. Companies with complex hybrid or multicloud environments may also bring in cloud architects or consultants to help with the design.
Map the network: Experts advise setting aside enough IP address space for future workloads, accounting for IPv6 and dividing the environment into subnets. Decide which resources need internet access, which need private connections to cloud services and which must reach other VPCs or on-premises systems. Then configure routing, gateways, private endpoints, VPNs and dedicated links around those needs.
Build in security and visibility: That same guidance extends to least-privilege policies, firewall and identity controls, encryption, flow logs and monitoring, followed by connectivity and failover testing before workloads move into production. Infrastructure-as-code templates and standardized policies can make deployments easier to repeat and manage as the environment grows.
Where Are VPCs Headed?
While VPC definitions continue to change, hyperscaler investment suggests the technology will be around for a while. At the same time, VPCs appear to be shifting from something IT teams configure individually to a more foundational building block.
AWS, for example, is using VPC Lattice to connect applications and services across multiple VPCs. Google is pushing service-centric networking through Private Service Connect and Cross-Cloud Network. And Microsoft is centralizing connectivity and security management across groups of virtual networks.
Agentic AI could influence this, too. Because agents act rather than simply retrieving information, they need access to assets and services across clouds and data centers, requiring their own private connections, centralized policies and networking.
Cloud networks are evolving into self-healing, automated foundations managed continuously by AI, according to John Capobianco, head of AI and developer relations at Itential. In his post What Is Agentic NetOps?, he explained how AI agents can dynamically plan, provision and self-correct VPC network configurations, security groups and routing tables in closed-loop automated environments.
"Agentic NetOps is…software that takes an operational goal, translates it into a plan, and executes that plan under governance,” he wrote.
In a podcast interview, Capobiano said he expects that by 2030, 70–80% of network infrastructure will be managed by AI agents. This moves VPCs deeper toward self-configuration.
Related:
This is an update to the original article published on January 25, 2023 and revised on September 17, 2024.
David Rand is a business and technology reporter whose work has appeared in major publications around the world. He specializes in spotting and digging into what’s coming next–and helping executives in organizations of all sizes know what to do about it.
Dipti Parmar and Joey Held contributed to earlier versions of this story.
© 2026 Nutanix, Inc. All rights reserved. For additional information and important legal disclaimers, please go here.
Related Articles
