Since VDI cost $500 per user a few years ago, it should be $300 today, right? Wrong!

One of the most hotly debated and discussed topics around VDI is cost, specifically, people wonder whether VDI is cheaper than traditional laptops and desktops.

One of the most hotly debated and discussed topics around VDI is cost, specifically, people wonder whether VDI is cheaper than traditional laptops and desktops. I've written about this quite a bit, including:

And of course:

That said, even in 2013 I still spend a fair amount of time discussing the cost of implementing VDI (CapEx) and how that compares to traditional PCs. The historic opinion in the industry has been that VDI isn't any cheaper to buy, but it's cheaper to maintain. (Though that's debatable, but not the point of this article.)

Often times when I mention that today, people respond with something like, "Hey, since VDI cost $500 per user a few years ago, it should be $300 per user today, right?"


The main reason they think that is because of Moore's law. Even though Moore's law technically applies to transistor density on microchips, it more typically manifests itself in our lives in the form of technology getting cheaper year after year. (Or, perhaps more commonly, the same amount of money being able to purchase increasingly faster and better gadgets.)

So when it comes to VDI, people think that Moore's law should be driving down the price. But that's not actually true. (At least not yet.)

In 2013, VDI is still going to cost you $500 per user. What Moore's Law has done has made it so you can buy a lot more for your $500 today versus back in 2009, (when Windows 7 was released). The advantage that we have is that since most people are building their VDI around Windows 7, the system requirements for a good experience are sort of "frozen" at 2009 levels. For example, Microsoft says the Windows 7 system requirements are:

  • 1GHz processor
  • 1GB RAM
  • 16GB disk space
  • DirectX 9 graphics device with WDDM 1.0 driver
Back in 2009, it cost $500 to buy a business-class PC with those specs. But looking at 2013's $500 PCs, you can get a Core i3 with 4GB RAM:



The same applies to VDI. Moore's Law means that the focus in 2013 is not that you can buy VDI for 1/5th the cost of VDI in 2009, but instead that the $500 you spend for VDI in 2013 is 5x more powerful per user than what you could buy in 2009. In other words the same money buys a much better experience.

This is huge—something that I probably should have included as Item #3 in my March 2013 article "2013 is the year that two of the biggest showstoppers to VDI adoption are finally solved." Five years ago we were really scraping the bottom of the barrel when it came to the amount of iron we could throw at a VDI desktop for $500. But in 2013, thanks to Moore's Law and the economies of scale we get in the datacenter and the fact that Windows 7's system requirements are not moving, we can actually put together a very capable VDI desktop for $500 per user. Huzzah!

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

Also don't forget that most of the desktop virtualization (VDI) products today offer builtin clever caching (vWorkspace HyperCache / VMware CBRC) which improves performance even more!


Good points Brian

From what I've been seeing in my market over the last 2 years or so, people are over the "cheaper than a pc " play.

The VDI projects that have kicked off are all around a business case like security, company acquisitions, or bizarrely 6 Star Energy rated buildings. The last one is coming up more commonly as the power load of a desktop PC is banned, buts it's ok to have a Thin Client, as long as the VDI server architecture lives in a different, non 6 Star Rated building :-)

Funny I know, but hey that's what we are seeing Down Under.

Catch ya



I've been around since WinFrame days.  We always thought/hoped I guess that Citrix access would be everywhere and that we'd finally rip out those PCs from  people's desks.  However things along the way like instability of Citrix with printing, less than optimal user experience vs traditional PCs, has meant that Citrix wasn't the widespread success as expected.

With dropping PCs prices and in some cases Wyse terminals being quite expensive, customers just haven't seen the value as rolling out nice and cheap PCs and "managing them" with traditional tools.  Customers are then not locked into VDI or hardware terminal vendors.  They are locked enough into Microsoft !   PCs are dirt cheap and they can roll them over when they fail.  With PCs now having flash drives, they will now last even longer.

Now users are coming along with nice shiny Apple MacBooks so again the value of having VDI has diminished even more and it will continue to be a point/tactical solution for some specific use cases.

With more and more web based applications and applications written for mobile devices, there is less pressure on the business to offer Windows access to key business applications.  This again devalues VDI for organisations.

I love the idea of centralising Windows apps and Citrix/VDI are great technologies.  I simply don't see them gaining widespread adoption in most companies for the foreseeable future.  I think we can continue to joke about the "this is the year for VDI" for a few more years to come.


Too bad VDI licensing doesn't obey Moore's law.


I doubt that many people today would try to deliver Win7 using the specs offered: 1GHZ CPU and 1GB RAM. I know I wouldn't want to live with that, and it is much less than what the vendors recommend.

Sounds like a typical low end notebook. We know how succesful those were in the marketplace, and the poor quality of user experience they delivered.

Another factor to consider, beyond the cost of the Microsft licencing as mentioned by "Rentvent", is the changing nature of desktop workloads and the ever changing ways we work.

Today, we use far more HD graphics, video and VoIP than we did just four years ago. To accomodate these applications means that the cost of delivery (Nvidia GPU's, upgraded network, higher perofromance SANs etc) goes up significantly. And deployment gets exponentially more complex.

And workers today are increasingly more mobile. This means that we are also required to deploy, support and pay for a second device, typically a laptop. And at a minimum, that also means more Msoft licence expense, at least if you wish to be fully compliant (Sarbox anyone?).

And let's not forget the drive to browser and mobile based apps, even Mac apps. All of these make the question "Is VDI Necessary?" all the more difficult to answer, and cost and complexity more dificult to justify.

The vaue prop for VDI should not be cost based. I would suggest that has pretty much been proven false. And as Brian once said, "if you want a $300 desktop, just buy a $300 desktop".


Wow! $500/user/yr, i.e. $40/user/month. Time to give 'Mobile Virtualization' from Armor5 a try, You get access to all your productivity apps a whole lot cheaper, and moreover, it's mobile form factor friendly, touch enabled and highly network/CPU efficient.

It's a whole new way of virtualizing applications and data using power of HTML5.



Your answer is actually absolutely correct - the cost per VDI user, hardware wise, is not $300, with proper architecture it is ... under $200, while offering linear scalability and excellent performace. Check this out:

Will be glad to discuss with you MicroPOD architecture that we were working on for the past 4.5 years. And yes, we do have real VDI users - over 5,000 in production, about 10,000 are on the way as part of several ongoing projects. And no, we are not using SSDs.

PS. In 2013 the industry average hardware cost per VDI user appears to be at about $800-1,000, or even higher, not even accounting for the cost of the endpoint.

Best regards,