What Is ESXi?

Most enterprise data centers no longer run one application per server. Instead, a single physical machine hosts dozens of separate workloads at once, each behaving as if it owns the hardware. The software that makes this possible is a hypervisor, and in VMware environments, that hypervisor is ESXi.

VMware ESXi is a Type 1, or bare-metal, hypervisor that installs directly onto a physical server's hardware—with no host operating system underneath—and divides that hardware into multiple isolated virtual machines (VMs), each running its own operating system and applications. It sits at the foundation of the broader VMware vSphere platform.

ESXi is one of the most widely deployed hypervisors in enterprise IT, yet it's often confused with vSphere and vCenter, the products built around it. The distinction matters when you're planning, sizing, or budgeting a virtualization project.

This article explains what ESXi is, how it works, how it differs from the rest of the VMware stack, and the two factors that shape real-world performance the most: the storage beneath it and the licensing model around it.

What is ESXi?

ESXi is a bare-metal hypervisor, meaning it runs directly on server hardware rather than as an application inside Windows or Linux. Because it installs straight onto the physical machine, a Type 1 hypervisor loads before any general-purpose operating system and takes direct control of the server's CPU, memory, storage, and network resources.

That direct hardware access is what separates ESXi from desktop virtualization tools. There's no host OS competing for resources or adding a layer of vulnerability between the hypervisor and the hardware.

At the core of ESXi is the VMkernel, a purpose-built kernel that schedules hardware resources and runs the virtual machines. ESXi's VMkernel interfaces directly with VMware agents and approved third-party modules, which keeps the footprint small and the attack surface narrow.

The name ESXi descends from VMware's original hypervisor, ESX, which stood for Elastic Sky X. The "i" signals the integrated version that replaced it.

ESXi serves enterprises, smaller organizations, and individual practitioners alike, but its real strength shows in production data centers where consolidation and reliability are non-negotiable.

How does ESXi work?

ESXi creates a virtualization layer between the physical hardware and the virtual machines running on it. When you install ESXi on a server, it claims the underlying CPU, memory, storage, and network adapters, then presents virtual versions of those resources to each VM.

Every virtual machine receives its own virtual hardware: virtual CPUs, memory, a virtual disk, and virtual network interfaces. From inside the guest operating system, that VM appears to run on dedicated hardware, with no awareness of the other VMs sharing the same host.

This isolation is a defining feature. Each VM is walled off from its neighbors, so a crash, misconfiguration, or compromise in one virtual machine doesn't cascade into the others on the same server.

The VMkernel schedules access to physical resources, enforces the boundaries between VMs, and translates requests between virtual and physical hardware. Because ESXi runs on bare metal rather than on top of an operating system, VMs achieve near-native performance, and the host can support large consolidation ratios.

A practical example: A single ESXi host might run a database server, a web application, an internal file share, and a test environment side by side. Each gets a defined slice of resources, and the administrator can adjust those allocations as demand shifts—often without taking the others offline.

ESXi vs. vSphere vs. vCenter

These three terms get used interchangeably, but they describe different things. Sorting them out is the most common point of confusion for teams new to VMware.

• ESXi is the hypervisor, the engine that runs virtual machines on a physical host.
• vCenter Server is the centralized management layer that controls many ESXi hosts from one console.
• vSphere is the full product suite that combines ESXi, vCenter, and advanced features such as vMotion and High Availability into one platform.

Put simply, ESXi is where virtual machines live, vCenter is where you manage them at scale, and vSphere is the umbrella that ties everything together. You can run a single ESXi host on its own, but once you operate more than one, vCenter becomes the tool that makes the environment manageable.

The features that most people associate with VMware—live migration, automatic failover, resource balancing—come from vSphere and depend on vCenter coordinating multiple ESXi hosts. ESXi supplies the raw virtualization; the surrounding suite turns it into a resilient platform.

From ESX to ESXi: How the hypervisor evolved

ESXi's architecture is the result of a deliberate simplification. The original ESX included a Linux-based Service Console for management, which added size, complexity, and potential vulnerabilities.

ESXi removed the Service Console entirely. The result is a hypervisor with a much smaller footprint—small enough to install from a USB device or SD card—and a reduced attack surface. VMware completed the full transition with vSphere 5.0 in 2011, after which ESXi became the only hypervisor the platform supported.

The version history tracks the platform's maturation:

• ESX 3.x (2006): The original hypervisor with the Linux-based Service Console
• ESXi 4.x (2009): First widespread adoption of the ESXi architecture
• ESXi 5.x (2011): Full transition away from ESX; the Service Console is gone
• ESXi 6.5/6.7 (2016–2018): Improved web client, VM encryption, and Secure Boot
• ESXi 7.x (2020): Kubernetes support and lifecycle management enhancements
• ESXi 8.x (2022): Support for data processing units (DPUs) and AI/ML workloads

Each release matters for planning because VMware ties support windows to versions. VMware typically supports ESXi for around seven years—roughly five years of general support followed by two years of technical guidance—so running an unsupported version carries real security and compliance risk.

Key features of ESXi

ESXi packs enterprise capability into a compact package. Its core features explain why it remains a default choice for production virtualization.

• Small, efficient footprint: The bare-metal design uses minimal resources for itself, leaving more of the hardware available to virtual machines.
• Scalable capacity: A single VM on ESXi in vSphere 8 can be configured with up to 768 virtual CPUs and 24TB of RAM, far beyond what most workloads require.
• Security controls: VM encryption, role-based access control, Secure Boot, and TPM 2.0 support protect both the hosts and the workloads running on them.
• Advanced platform features: Through vSphere, ESXi hosts gain vMotion for live VM migration, High Availability for automatic restart after a host failure, and the Distributed Resource Scheduler (DRS) for balancing workloads across a cluster.
• Flexible management: Administrators manage hosts through the vSphere client, command-line tools, or REST APIs, depending on whether they're working interactively or automating at scale.

These features rarely operate in isolation. Encryption, role-based access, and detailed logging work together to give administrators both security and an audit trail for troubleshooting.

Business benefits of ESXi

The technical capabilities translate into measurable operational and financial outcomes, which is why virtualization became standard practice in the first place.

• Server consolidation and lower costs. Running multiple VMs on one host increases hardware utilization and shrinks the physical server count. Fewer machines mean lower spending on hardware, power, cooling, and data center space. For example, a company that consolidates 15 aging servers onto three modern hosts cuts not just acquisition cost but also ongoing energy and maintenance overhead.
• Higher availability. Paired with vSphere, ESXi supports failover and live migration. If a host fails, workloads restart on a healthy host automatically. If a host needs maintenance, vMotion moves running VMs elsewhere with no noticeable downtime.
• Faster provisioning. Spinning up a new server no longer means racking hardware. Administrators deploy VMs from templates in minutes, which speeds testing, development, and scaling.
• Cleaner separation of environments. Because VMs are isolated, teams can run test and production workloads on the same hardware without one interfering with the other.

How storage shapes ESXi performance

Here's what many ESXi explanations skip: The hypervisor’s performance and resilience depend on the storage beneath it. ESXi keeps every virtual machine's files—its virtual disks, configuration, and snapshots—on a storage location called a datastore. Where that datastore lives largely determines how the environment performs and which features you can actually use.

A datastore can sit on local disks inside the host, or on shared external storage that multiple hosts can reach over a network using protocols like Fibre Channel, iSCSI, or NVMe over Fabrics. The choice has consequences beyond raw speed.

• Shared storage unlocks the cluster features. vMotion, High Availability, and DRS all depend on multiple ESXi hosts seeing the same VM files. With local storage, a VM is trapped on the host that owns its disk. With shared external storage, any host in the cluster can run any VM, which is what makes live migration and automatic failover possible. Teams that build clusters on local disk often discover too late that the features they bought vSphere for won't work.
• Datastore format affects management granularity. ESXi supports its native VMFS file system as well as Virtual Volumes (vVols), which map storage operations to individual VMs rather than to a shared volume. vVols let storage-level snapshots and policies apply per virtual machine, which simplifies protection and performance tuning.
• Storage latency is often the real bottleneck. When VMs feel slow, the cause is frequently storage I/O rather than the hypervisor or CPU. Flash-based arrays and low-latency transports like NVMe over Fabrics can keep response times low even as consolidation ratios climb, which is why storage performance deserves as much planning attention as the hosts themselves.

What ESXi licensing looks like under Broadcom

ESXi's licensing has changed more in the past two years than in the previous two decades, and it now factors heavily into virtualization strategy.

Broadcom completed its acquisition of VMware in late 2023. Shortly after, in early 2024, VMware ended the sale of perpetual licenses across its portfolio. Customers can continue using licenses they already own, but new purchases and ongoing support now run through subscriptions priced per core.

The product catalog has also been consolidated. Standalone editions have given way to bundled subscriptions, primarily vSphere Foundation (VVF) for organizations that need reliable virtualization and management, and VMware Cloud Foundation (VCF) for a fuller private-cloud platform that adds storage, networking, and automation. The free standalone hypervisor and several entry-level kits have also been affected as part of this consolidation.

For many IT teams, these shifts have reopened a question that seemed settled: stay on VMware or evaluate alternatives. As organizations weigh that decision, one practical point often gets overlooked—The storage layer is portable. External storage that serves an ESXi environment today can serve a different hypervisor tomorrow, which means a storage investment doesn't have to lock you into a single virtualization platform. Separating the storage decision from the hypervisor decision gives teams more room to adapt.

Best practices for running ESXi

A few decisions early in an ESXi deployment can help prevent the most common problems later.

1. Verify hardware compatibility first. ESXi runs on a defined set of supported hardware. Confirm your server, network adapters, and storage are on the compatibility list, and enable hardware virtualization in the system BIOS before installing.
2. Use shared storage for any cluster. If you plan to run more than one host and want vMotion, HA, or DRS, build on shared external storage from the start. Retrofitting it later is disruptive.
3. Build in redundancy. Use redundant boot media and resilient storage configurations so a single drive failure doesn't take down a host.
4. Right-size virtual machines. Overallocating vCPUs and memory wastes capacity and can hurt performance. Match resources to actual workload demand, and watch for VM sprawl—forgotten VMs that consume resources and widen the attack surface.
5. Harden security. Apply role-based access control, disable services you don't use, keep hosts patched, and take advantage of Secure Boot and TPM where available.
6. Protect VM data against ransomware. Snapshots alone aren't a backup. Keep immutable, off-site copies of critical VMs so you can recover even if production storage is encrypted or deleted in an attack.

The last point deserves emphasis. Virtualization concentrates many workloads onto shared infrastructure, which means a single ransomware event can affect dozens of VMs at once. Recovery planning needs to occur in the deployment phase, not after an incident.

Conclusion

VMware ESXi is the bare-metal hypervisor at the foundation of the VMware virtualization stack, the layer that turns a physical server into a pool of isolated virtual machines. Understanding how it relates to vSphere and vCenter, and recognizing that its performance and resilience depend heavily on the storage beneath it, separates a smoothly running environment from a fragile one.

For most organizations, the decision to virtualize was made long ago. The decisions that matter now are about how to run ESXi well: sizing it correctly, building on storage that enables the features you're paying for, and protecting the workloads it hosts. Those choices can determine whether virtualization delivers the consolidation, uptime, and agility it promises.

That's where the storage layer earns its keep. Everpure™ FlashArray™ provides external storage for vSphere environments that supports both VMFS and Virtual Volumes, delivering the low-latency performance and shared access that vMotion, High Availability, and DRS require. To guard virtualized workloads against ransomware and accidental deletion, SafeMode™ Snapshots create immutable copies that attackers cannot alter or erase, supporting rapid recovery of affected VMs. For teams that prefer to consume storage as a service rather than manage hardware refresh cycles, Evergreen//One™ delivers capacity on demand with the flexibility to scale as virtualization workloads grow. 

Whether you stay on ESXi or evaluate alternatives, a storage foundation built for performance and recoverability helps keep your options open.