With Citrix pricing models compared side by side, the same environment can carry wildly different costs before a single discount is negotiated. User, device, and concurrent licensing each measure consumption in a different way, and the model you sign under decides the quantity the vendor prices. Get the model wrong and you pay for capacity you never use. Get it right and you remove shelfware that no rate reduction could match. This is why the model decision is the first and often the most valuable lever in any Citrix deal, and why it should be made on your own usage evidence rather than the vendor's recommendation. This guide explains how the three models work, where each one wins, and how to turn the choice into negotiation leverage.

Unsure which model fits your usage? The wrong licensing model can cost more than a bad rate. Contact us for a free review of your Citrix usage and the model that actually matches it.

Citrix pricing models compared: how user, device, and concurrent licensing differ

The three models answer the same question, how much access are you buying, in three different units. User based licensing assigns an entitlement to each named individual who is permitted to access the environment, whether or not they log in on a given day. Device based licensing ties the entitlement to a specific endpoint, so any number of people using that one machine consume a single license. Concurrent licensing ignores the total population entirely and counts only the peak number of simultaneous sessions, so a pool of licenses serves a much larger headcount as long as they do not all connect at once. The distinction matters because real organisations almost never have a usage pattern where all three models produce the same number. Understanding which unit the vendor is counting is the starting point, and it sits inside the wider set of levers mapped in our Citrix negotiations guide.

Where named user licensing wins

User based licensing is the cleanest match when each person needs persistent, individual access and the population logs in regularly. Knowledge workers with assigned desktops, remote staff who connect daily, and any group where the headcount and the active user count are close together all suit a named user model, because there is little gap between people entitled and people using. The trap appears when an organisation defaults to named user out of habit while its real pattern is intermittent. If half your named users connect only occasionally, you are funding entitlements that sit idle, and a concurrent model would price the same access for far less. Named user is the right answer often enough that vendors propose it readily, which is exactly why it deserves scrutiny against your measured usage rather than acceptance as the obvious default.

The model you sign under decides the quantity the vendor prices, which is why it often matters more than the rate.

Where device licensing wins

Device based licensing comes into its own wherever a single endpoint is shared by many people across a day. Shop floor terminals, hospital ward workstations, call center desks running multiple shifts, and kiosks all fit, because the device is constant while the people rotate. In these settings a device license can cover three or four shift workers on one entitlement, which a per person model would charge for individually. The decision hinges on the ratio of people to devices: where that ratio is high and devices are genuinely shared, device licensing is frequently the cheapest compliant position. Where most people have their own machine, the model offers no advantage and can cost more. As with every model, the answer comes from counting your real device and user populations, not from assuming shared working is widespread when it may not be.

Where concurrent licensing wins

Concurrent licensing is the model that most often surprises buyers with savings, because it prices the peak rather than the population. Any environment where the total headcount substantially exceeds the number of people connected at the busiest moment is a candidate. Shift based operations, seasonal businesses, task workers who log in briefly, and large user bases with staggered working hours all tend to have concurrency peaks far below their nominal user count. When that gap is wide, a concurrent pool sized to the true peak can serve thousands of people for a fraction of the per user cost. The catch is that concurrent licensing rewards accurate measurement and punishes guesswork: size the pool too low and you risk denied sessions at peak, too high and you give back the saving. This is why a measured concurrency curve, the same discipline behind our licensing advisory service, is the foundation of any concurrent model decision.

The 2026 packaging context

The model conversation does not happen in a vacuum, and the packaging landscape has shifted. As of June 2026 Citrix sells under Cloud Software Group on a subscription only basis, perpetual licensing having ended in October 2022, and current commercial packaging centers on the Citrix Platform license and Universal Hybrid Multi Cloud licensing alongside the established user, device, and concurrent constructs. The practical effect is that the clean comparison between three classic models is increasingly mediated by which package the vendor proposes, and packaging can bundle or obscure the underlying unit. Always confirm in writing which model actually governs the entitlements you are buying, and frame any pricing or availability point with an as of reference, because packaging and the models available within it continue to change. The negotiation principles behind the Platform license are covered in our guide to negotiating Citrix Platform license pricing.

Turning the model choice into leverage

The reason this matters commercially is that the model sets the baseline quantity before any discount is even discussed. A vendor that selects the model maximising counted units starts you from an inflated number, and a percentage discount off an inflated number is worth less than the right model applied to a smaller one. The disciplined sequence is to measure first, then choose the model that matches your usage, then negotiate rate on the defensible quantity that remains. Doing it in that order removes shelfware that no rate concession could touch, and it gives you an evidence backed reason to reject a model that does not fit. Where a model switch lowers the vendor's counted units, expect resistance, and bring a credible alternative to the table to hold your position. The mechanics of building that pressure are set out in our guide to competitive alternatives as negotiation leverage.

Choosing the right model for your estate

There is no universally cheapest Citrix pricing model, only the one that fits your usage, and the only way to find it is to put your real numbers against each option. Measure your named population, your device sharing ratios, and above all your true concurrency peak, then price user, device, and concurrent against the same estate and let the evidence decide. Treat the vendor's proposed model as one option to be tested, not the answer, and remember that switching models at renewal is a negotiated outcome you win with data, not a menu the vendor offers freely. We are independent Citrix licensing experts, 100 percent buyer side, with no reseller or vendor affiliations, and our senior advisors have vendor side backgrounds, so we know how packaging is used to steer buyers toward the unit that suits the vendor. The full service sits on our Citrix negotiation service page.

Frequently asked questions

What are the main Citrix pricing models compared?

The main Citrix pricing models compared are user based, device based, and concurrent. User based licensing assigns an entitlement to each named person who may access the environment. Device based licensing ties the entitlement to a specific endpoint regardless of who uses it. Concurrent licensing counts the maximum number of simultaneous sessions rather than the total population. Each measures consumption differently, so the same estate can carry very different costs depending on which model applies.

Which Citrix licensing model is cheapest?

There is no single cheapest model. The cheapest model is the one that matches your actual usage pattern. Concurrent licensing tends to win when many people share a smaller pool of simultaneous sessions, such as shift workers or task workers. User licensing tends to win when each person needs persistent individual access. Device licensing can win in shared kiosk or shop floor settings. The only way to know is to measure your real concurrency curve and population, then price each model against it.

Does Citrix still offer concurrent licensing in 2026?

Availability of specific models changes with packaging, and as of June 2026 Citrix sells under Cloud Software Group on a subscription only basis following the October 2022 end of perpetual licensing. Concurrent style entitlements have historically been part of CVAD packaging, but current commercial packaging centers on the Citrix Platform license and Universal Hybrid Multi Cloud licensing. Confirm what is actually on offer for your renewal in writing rather than assuming a model is available, because the vendor steers buyers toward the packaging that suits its revenue, not yours.

How does the pricing model choice affect a Citrix negotiation?

The model choice sets the quantity being priced before any discount is discussed, so it often matters more than the rate. A buyer who lets the vendor pick the model that maximises counted units starts from an inflated baseline. A buyer who measures usage and selects the model that matches it removes shelfware first, then negotiates rate on a smaller, defensible number. The model decision is leverage, and it should be made on your evidence, not the vendor's recommendation.

Can you switch Citrix pricing models at renewal?

Switching models at renewal is sometimes possible but depends on current packaging and what the vendor will agree to. A move from named user to concurrent, or into a Platform license, can produce real savings when usage supports it, but the vendor may resist a change that lowers counted units. Treat a model switch as a negotiated outcome backed by usage evidence and, where useful, a credible alternative, not as a menu choice the vendor offers freely.