Versa Networks Is Selling Less Network - and That’s the Point

The least interesting object in a modern branch office may be the most expensive: a beige or black box blinking quietly in a rack. One routes traffic. Another filters it. A third inspects threats. Someone added a cloud-security service during the remote-work scramble, and a fourth dashboard now watches the other three. Each product is defensible. Together they form a small government.

Versa Networks wants to dissolve that government. The privately held company from Santa Clara sells VersaONE, a software architecture that combines wide-area networking, cloud-delivered security, local-area networking, firewalls, analytics and policy. It can run in Versa appliances, in public clouds, on premises or in some mixture of the three. The practical promise is not mystical: one console, one policy model, one pool of telemetry and fewer seams for attackers or operators to fall through.

Before SASE had a name

Versa began in 2012, seven years before Gartner coined the category name Secure Access Service Edge, or SASE. The founders, brothers Kumar and Apurva Mehta, had helped build Juniper Networks’ MX router line. Their original complaint was aimed at hardware rigidity. Networks could not change as quickly as the applications using them, and service providers needed weeks or months to roll out functions that software should have made programmable.

So the brothers moved the functions into software. Early Versa language centered on virtual network functions: routing, security and traffic management that could run on commodity machines rather than a dedicated box for every job. The terminology aged badly. The architecture aged well. When cloud applications, remote work and zero-trust security pushed the network perimeter into hundreds of locations, the market found a new acronym for much of what Versa had already assembled.

Networks needed programmability and agility to allow them to adapt as quickly as applications.Kumar Mehta, explaining the company’s origin

What exactly did Versa do? It put security and networking into the same operating system and designed the traffic path so several jobs could happen in one pass. A packet can be identified, routed, inspected and governed without bouncing through a chain of separately managed services. The software is genuinely multi-tenant, an unromantic feature that matters enormously to a telecom carrier or managed-service provider running the product for many customers.

The bill that changed the mind

Architecture diagrams are patient. Corporate budgets are not. Adobe’s move to Versa supplies the cleanest public example of what broke first and what changed the buyer’s mind. Its legacy wide-area network leaned on costly private circuits. Cloud traffic demanded more bandwidth, employees expected better application performance and management wanted lower operating costs. The old model could satisfy two of those demands, but not all three.

Adobe reported that its previous network cost a little over $500,000 per site across five years. The Versa SD-WAN model came to a little over $200,000 in the company’s comparison, or roughly 50 percent less over five years. Moving primary traffic from MPLS to internet circuits created the large economic change. Versa supplied application-aware routing, monitoring and security so that less expensive transport did not have to mean blind transport.

The first failure was not a router. It was understanding. The new fabric changed how engineers operated the network, and the project team found that a technical rollout needed an internal marketing campaign. It set up education classes and hands-on labs, ran lunch-and-learns and took the vision on a speaking tour across the company. The network was programmable; the organization still needed persuasion.

We did a lot of our own marketing, like lunch-and-learn events to educate across the business.An Adobe team member on the rollout

One platform, four doors

VersaONE is best understood as an umbrella with four useful entrances. Secure SD-WAN chooses paths for application traffic and keeps branch connectivity alive when a link fails. Security Service Edge applies zero-trust access, web filtering, cloud-app controls and data-loss rules to users wherever they work. Secure SD-LAN extends policy inside campuses and branches. Advanced security supplies firewalls, intrusion prevention, malware defense and controls for connected devices.

VersaAI now sits across those doors. One half uses machine learning to spot anomalies, diagnose performance and automate operations. The other half protects enterprise use of generative AI and the growing population of nonhuman identities. The useful raw material is not the label “AI.” It is the combined network and security telemetry underneath it. A model can draw better conclusions when user identity, device posture, route performance and suspicious traffic live in the same record.

Where Versa fits in a crowded market

The alternatives are formidable. Cisco and Fortinet approach the problem from networking and appliances. Palo Alto Networks, Zscaler and Netskope approach it from security. Cato Networks built a cloud-first network and security service. Versa’s position between those camps is specific: carrier-grade routing, a broad security stack, real multi-tenancy and the freedom to run in the cloud, on premises or both.

That deployment choice is more than procurement decoration. A global manufacturer may need security at a plant where cloud connectivity is unreliable. A bank may want local inspection for sensitive traffic. A European public body may require the data plane, control plane and management plane to stay inside a jurisdiction. Versa’s 2026 European sovereign service turns those constraints into a product rather than an exception ticket.

The company was named a Challenger in Gartner’s 2026 SASE Platforms report, its fourth consecutive appearance. It says its technology reaches more than 20,000 organizations in over 100 countries. Those customers arrive both directly and through carriers, managed-service providers, resellers and distributors. In 2021, Versa counted 150 service providers, including seven of the world’s ten largest. That route to market is part of the product: multi-tenancy gives partners a way to package secure networking as a managed service.

The business is consolidation

Versa earns money from software licenses, cloud subscriptions, support and services sold to enterprises and through the channel. Its CSG and CSX appliances are purchased through distributors, not directly from Versa. The company has raised approximately $316.25 million across rounds that began with $14.35 million in 2012 and culminated in a $120 million BlackRock-led financing in 2022. It called that last deal a pre-IPO round. Four years later, it remains private.

The patient capital paid for breadth. A useful SASE product cannot stop at drawing tunnels between branches. It must recognize applications, enforce identity, inspect encrypted traffic, protect data, report user experience and work across clouds. Versa chose to build or integrate all of that around one architecture. The result can remove licenses and consoles. It can also make the buying decision feel like choosing an operating system for a large part of the company.

When the pitch does not work

Consolidation is not automatically cheaper, safer or easier. A small company with few sites may not have enough sprawl to justify a platform migration. A security team may prefer a specialist cloud-security service while the network team keeps a favored SD-WAN product. A regulated organization may require controls the unified stack has not yet matched. And an enterprise with scarce network talent can buy one console only to discover that the console still contains a great deal of network.

This is why the Adobe story is more useful than a feature checklist. The savings came from changing transport economics, not merely installing software. The operational win came after a campaign of classes, labs and conversations. A buyer can copy that sequence: measure the expensive seam, pilot the new architecture on representative sites, teach the operators, publish the results and migrate in waves.

The harder lesson is architectural. Versa started with a seam between networking and security and refused to treat integration as an afterthought. Other companies can copy the method: find two functions whose separate data and workflows create visible expense, then build them around a shared control plane. What they cannot copy quickly is the Mehta brothers’ decades of routing experience, a service-provider channel or fourteen years of failure testing.

Versa’s most credible promise is therefore modest. It will not make a global network simple. Global networks are collections of geography, regulation, old equipment and human judgment. It can make the complexity live in fewer places. For an infrastructure team juggling a cabinet full of boxes and a browser full of consoles, fewer places may be plenty.