Will it work?

Running Revit in the Cloud: 10 Factors That Determine Performance

Is Revit in the Cloud as fast as your workstation, or unbearably slow ? In a recent poll of over 500 Revit users, we found strong interest in running Revit in the cloud—but also concerns about performance. Will input lag slow down workflows? Can large models sync efficiently? Does internet stability impact productivity? This article breaks down the 10 key factors that determine Revit’s performance in the cloud, covering hardware, network, and collaboration setups. More importantly, we show you how to test and optimize each factor yourself. If you're considering a move to the cloud, this guide will help you make an informed decision. Read on to see if Revit in the cloud can match (or even beat) your workstation.

February 18, 2025
Don Rekko
Don Rekko
Running Revit in the Cloud: 10 Factors That Determine Performance

In a recent Designair poll with over 500 Revit users, we discovered  a strong interest in running Revit in the cloud. Many users see the potential advantages: flexibility, access from anywhere, and centralized management. 

However, some of our survey respondents wondered how cloud-based workstations would perform compared to their high-end local machines, particularly when working with large models. 

In this article, we’ll dig into more detail about performance issues related to running Revit in the cloud - and how they can be managed. 

Performance concerns about running Revit in the cloud

In the survey, latency and performance issues were among the top concerns for users considering running Revit in the cloud. The concerns generally fell into three key categories:

1. Input Lag and Real-Time Interactions

  • Many users worry about input delay affecting modeling, drafting, and rendering workflows.
  • Even slight latency can impact efficiency and user experience, making fast and responsive cloud setups critical.

Here’s what one respondent said: “...I have concerns about potential input lag when running Revit entirely in the cloud. Given the real-time nature of modeling, drafting, and rendering workflows, any latency in user input could impact efficiency and the overall user experience.”

2. Internet Reliability & Stability

  • Cloud-based Revit is completely dependent on a stable internet connection.
  • Internet outages can cause costly downtime, leaving entire teams unable to work.
  • Users in construction sites or remote areas are especially concerned about losing access.

On this topic, one respondent said that ”...[T]he only issue with a completely cloud-based solution is the total dependence on an internet connection. We have hosted models on the cloud since 2017, and to be fair, there have only been three outages in that time. But when you have 12 technicians staring at the wall for 8 hours, it costs a lot of money. At least with a local machine running Revit, you can switch to offline work and minimize loss.”

3. Performance Variability for Large Models

  • Users noted that file size affects performance—small projects may work fine, but large Revit models slow down significantly in the cloud.
  • Saving, syncing, and opening files from the cloud can sometimes be much slower than local storage
  • Some raised concerns around network-related delays when working on large, complex, multi-user models.

Another survey respondent put it like this: “It seems that placing models and such on the cloud slows saving, syncing, and opening models down—at times, considerably. I know as the technology has progressed, it's improved, but we all remember the bad experiences and forget every time it's worked flawlessly.”

This article will not try to debunk, deny or discredit these concerns. The aim is to break down these potential shortcomings in single factors, and discuss how each factor can impact your experience with running Revit in the cloud. We will show you how you can identify these factors, and how you can remedy and troubleshoot them. 

Nothing Beats a Hands-On Test

No matter how much research you do, the best way to evaluate a cloud workstation is to test it on your own device, from your own network, with your own version of Revit, while using your own workflows and data. This eliminates guesswork and ensures that the performance aligns with your specific needs. Every setup is different, and real-world testing is the best way to assess whether a cloud-based solution works for you.

With that said, for some that is a very unsatisfying approach - especially if that test confirms that Revit runs poorly in the cloud.

10 factors that affect the performance of Revit in the cloud

The performance of Revit in the cloud depends on 10 factors, either related to computing resources, streaming technology or network conditions. This article outlines these 10 factors that make or break a smooth performance of Revit in the cloud. As well as explaining what these factors are, we give you the tests and means to test each of them for yourself. 

These 10 factors are: 

Compute: 

  1. CPU Performance (clockspeed)
  2. RAM (Memory Size)
  3. GPU Type and size
  4. Disk

Streamer

  1. Streamer technology

Network

  1. Latency
  2. Bandwidth
  3. Stability and consistency
  4. VPN

Collaboration setup

  1. LAN-based, FileShare based or ACC/BIM360

Computing Resources: The Role of CPU, GPU, Memory, and Storage

In any environment—cloud or local—the performance of Revit is influenced by the underlying hardware. Here are the first four factors that shape the experience:

1 - CPU clockspeed

Revit is primarily single-threaded, meaning it benefits most from high clock speeds rather than a large number of cores. Certain workflows, such as rendering, can take advantage of multiple cores, but most everyday tasks rely on fast per-core performance

Here’s how you can test if the CPU of your cloud computer delivers the right performance when you work with Revit. 

Open a large model, switch between views, and see if there's a delay in regeneration. While you test this, do the following: 

  • Check Clock Speed: (Press Ctrl + Shift + Esc), go to the Performance tab, and monitor the CPU Usage while working in Revit. Check if the peaks (1) don’t “hit” the 100% utilization. 
  • Check Clock Speed: (Press Ctrl + Shift + Esc), go to the Performance tab, and monitor the CPU Usage while working in Revit. Check if the baseload (2) has some “headroom”. This is less important, but our recommendation is to check that the base level does not exceed 80% of CPU utilization. 

If the tests show that your CPU is maxed out, or gets consistently over 80%, then ask your cloud provider to upgrade to a CPU model with a higher clock speed, if the additional cost and the performance improvement is worth it. 

2 - Memory (RAM), in particular for large Revit models

Revit projects often involve large datasets, and available RAM plays a big role in how smoothly models load, edit, and render. Autodesk recommends 16GB as a minimum. Larger, more complex models make use of more RAM. 

What makes Revit a bit different from many other CAD and BIM applications is how it handles project files. They are compressed when saved, but when you open them, Revit uncompresses them into memory. This means the file expands and takes up a lot more space in RAM than its saved file size.

To check how much RAM you need for Revit, Autodesk’s “rule of 20” can be helpful. This rule says that Revit uses about 20 times the size of the RVT file in RAM. If your RVT file is 1 GB, Revit will need 1 GB × 20 = 20 GB of RAM to work smoothly.

That’s not the whole story. Linked files also matter. If your project has linked RVT, DWG, or IFC files, you need to add their expanded sizes too.

Bear in mind that other software needs RAM too – including your operating system, browsers, and other software. Always leave extra headroom for them (add 10% - 20% RAM on top). 

If Revit runs out of RAM it will start using the hard drive (page file) instead, which will slow everything down significantly. That’s why you always need to ensure that your cloud provider does not allocate RAM dynamically - i.e., give you less RAM when more of your colleagues are working - but always gives you a dedicated RAM capacity. If you see big differences in pricing between providers, this is the number #1 reason for these price differences. 

One of the benefits of cloud computing is that (some) cloud providers can scale up the size of your memory. This comes, of course, at additional cost, but it is a great option, especially if you only need more RAM temporarily. 

Here’s how you can determine if your cloud computer has enough RAM for Revit. 

  • Open the Windows Task Manager: In Task Manager → Performance → Memory, look at the Headroom and Available RAM.

There’s one final thing to consider. When your CPU is constantly maxed out even when it has plenty of CPU and GPU, it feels like your cloud computer is “choking”. This happens when the software is misconfigured and doesn’t use the GPU, and tries to offload rendering tasks to the  CPU (which is not very good at floating point operations). This is not common with Revit, but may occur with Solidworks or when the display settings of your streamer are incorrect. Definitely work with your cloud provider and software vendor to resolve this. 

3 - GPU Type and size

While Revit itself does not rely heavily on a GPU for modeling, users who work with Realistic View modes, linked files with high-resolution textures, or external rendering tools may notice differences in responsiveness depending on how GPU resources are allocated.

Revit will not need more than 4GB GPU, according to this recommendation from Autodesk. It states that for minimum, standard and large models, a 4GB GPU suffices. 

Here is how to find the make and model of the GPU of your cloud computer. In this screen, you can also evaluate how much Revit utilizes your GPU. 

  • Check GPU: (Press Ctrl + Shift + Esc), go to the Performance tab, and monitor the GPU Utilization while working in Revit
  • You will also find the make and model of your GPU in this tab, as well as the dedicated capacity of your GPU. 

Final reminder: if you use Revit in conjunction with rendering and visualization tools, then you will need to size your cloud computer’s GPU for these workloads. This will require a much larger GPU than Revit needs; at least 16GB or ideally 32 GB. Unsurprisingly, this comes at a higher cost.

4 - Storage Access and Load Times

The way files are stored and accessed affects how quickly models open and save. In cloud environments, storage access may be structured differently than on a local SSD, influencing loading times for projects with multiple linked files. 

A practical way to test your setup is to upload a large Revit model to your cloud computer, and open it in Revit on your cloud computer. Compare it with the load time on your local rig. As you do this test on your cloud computer, observe what happens in the performance monitor. 

Here’s how you can monitor the performance of your Disk while you’re opening up that large Revit model. 

  • Check Disk Usage: (Press Ctrl + Shift + Esc), go to the Performance tab, and monitor the Disk Usage while working in Revit

If the loading time on your cloud computer is not acceptable and/or you see that the disk is maxed out during loading, then there’s always the possibility that the cloud provider can upgrade the storage type, e.g. to an SDD with higher throughput capacity. Better performance comes at an additional cost.  

Note that this is not the full story. If you’re using the Revit Central Model file or a cloud-based solution such as Autodesk Construction cloud for collaboration, there are some additional things to consider. We’ll cover this later (see "10 - Shared model access")

2. Streaming technology

Unlike a local workstation, where Revit runs directly on the user’s machine, a cloud workstation streams its display to the user’s device. This means that every interaction in Revit—zooming, panning, clicking, selecting—is transmitted over the network before appearing on the user’s screen.

5 - Streaming technology

  • When a user moves their mouse, the input is sent to the cloud workstation.
  • The cloud workstation processes the command and updates the display.
  • The new image is then compressed and streamed back to the user’s screen in real time.

Streaming is the fundamental technology that makes it possible to work remotely on cloud workstations over a network. The streaming technology is also an important factor and influences how smooth, responsive, and high-quality the experience feels. Providers of streaming technology have their own protocols that determine how display data is compressed, transmitted, and rendered on the user’s screen.

Some commonly used streaming protocols for high-end visualization include:

  • Remote Desktop Protocol (RDP) – Standard (and free) for Windows-based remote access. Can be optimized for performance. 
  • PC-over-IP (PCoIP) – Designed for high-fidelity graphics. 
  • HDX (Citrix) – Optimized for virtual desktops, balances performance and image quality but may introduce noticeable lag in certain setups.
  • NVIDIA RTX Virtual Workstation (vWS) with NVENC – Uses hardware-based encoding for smoother video streaming, reducing CPU load.
  • Teradici CAS (Cloud Access Software) – Optimized for CAD and 3D applications, offering low-latency, high-fidelity streaming.
  • Nice DCV (AWS) – Designed for high-performance remote visualization, supports multi-monitor setups and high-resolution workflows.
  • Splashtop – A widely-used remote access solution known for its simplicity and ease of setup. It provides a solid user experience for high-end visualization, but it may introduce some latency and visual glitches on large datasets.

At this point, you may be wondering: why not just pick the fastest/best protocol? Alas, each of the technologies comes with a set of trade-offs between the following: 

  • Cost – All protocols (with the exception of RDP) have additional licensing fees for use in a business environment. 
  • Security – Given the sensitive nature of BIM, you will need fully encrypted connections, secure authentication, and compliance with industry standards, which some protocols handle better than others. It’s especially important to mention that using RDP without taking further security measures makes your system highly vulnerable to cyberattacks.  
  • Bandwidth Requirements – High-quality protocols like PCoIP and NVIDIA vWS require more bandwidth, which may not be ideal in low-bandwidth environments.
  • Multi-Monitor Support – Some solutions, like Nice DCV, support multiple 4K monitors, while others may struggle with high-resolution setups.
  • Ease of Deployment and Support – Certain protocols require extensive setup and IT management, while others offer a simpler out-of-the-box experience. Another critical aspect is how well the vendor’s support is. Splashtop, for instance, has a good reputation for customer support. 

You will probably not have much choice picking a streaming protocol. But it is good to be aware of the crucial technology that is part of the overall solution. Second, each of these protocols lets you tweak settings that can dramatically improve (or kill) your experience. If all else works and looks fine, and your experience is lacking, there's a high likelihood that the streamer is causing your troubles. We advise you to work with your Desktop-as-a-Service provider to fix this. 

3. Network Considerations: Connection, Stability, and Responsiveness

Because a cloud-based Revit session relies ultimately on constant data exchange, network conditions play a major role in performance - no matter how good your streaming technology is. 

6 - Latency 

When working in Revit, you expect real-time feedback as you zoom, pan, and edit. The time it takes for inputs to register—measured as latency— will vary on geographic distance, connection type, and network conditions.

An ideal value for latency is 15 milliseconds or less, while up to 50 milliseconds is acceptable. The best way to determine your latency is to use the built-in latency tool of your streamer. They all have the ability to measure and track your latency. If you’re familiar with the ping command, you can try that. However, since most cloud providers will disable pinging by default as a security measure, it’s likely that you’ll run into that limitation. 

What if your latency is more than 50 milliseconds? There are a few things you can do to improve your experience:  

  • Switch to a wired connection if you’re working on a desktop or laptop. This is faster and more stable than Wifi. If possible: avoid public Wifi. 
  • Adjust settings in your streamer (e.g. lower resolution, reduce frame rate)
  • Close background applications that also use your network (Dropbox, OneDrive, Zoom, etc.)
  • Ensure your Ethernet/Wi-Fi drivers are up to date 

Here are some more advanced solutions: 

  • Upgrade your local device, especially if you have an older or unpowered device. A tip to test if a different device can provide a better experience, is to use a (fairly recent) mobile device (ideally a tablet) and see what the difference is. 
  • See if there are other Internet Service Providers available who offer glass instead of copper connections (and more bandwidth, while you’re at it). 
  • Check if you’re working from the closest data center that your cloud provider offers. 

7 - Bandwidth 

Using Revit in the cloud relies on high-bandwidth display streaming - not that different from streaming high-resolution content to your TV.  Since you’re working in real-time (instead of downloading movies), you need not only low latency, but also enough bandwidth to ensure a sharp, clear view.  When you have a limited bandwidth (10 Mb/s or less) your streamer will try to compensate and introduce visual glitches, particularly when working with high-resolution (4K) content. 

A good test for your bandwidth  is provided by SpeedTest.

Run these tests both at your local device and on the cloud workstation to get a feel for your bandwidth. The test is not an exact science and your bandwidth will vary over time, but it is “good enough”. 

What if your bandwidth is not 10 Mb/s? There are a few things you can do to improve your experience (these will also help with latency):  

  • Adjust settings in your streamer (e.g. lower resolution, reduce frame rate)
  • Close background applications that also use your network (Dropbox, OneDrive, Zoom, etc.)
  • Ensure your Ethernet/Wi-Fi drivers are up to date 

Here are some more advanced solutions: 

  • See if there are other Internet Service Providers, who offer glass instead of copper connections (and more bandwidth, while you’re at it). 

8 - Network Stability and Consistency

A consistent connection is just as important as speed. Factors like network jitter (variations in response time) and packet loss (missing bits of data) can contribute to an experience that feels less predictable at times. 

To measure this, you can use the built-in tool of your streamer. Or, you can do the following, somewhat more labor-intensive and non-scientific test: 

  • Run Speed Test three to five times in a row
  • Note down the latency, as well as the up- and download speeds 
  • If ping varies significantly (e.g. 40 ms, then 15 ms, then 25 ms) you have jitter
  • If download (or upload) speeds vary wildly, then you have packet loss 

There are two solutions for an unstable or inconsistent connection:

  • Use a cable, or 
  • Changing network provider (since the network of your current provider is probably congested) 

Before you take this drastic step, try to connect to your mobile hotspot (via your phone), and see if the bandwidth greatly improves. If so, switch providers (or keep working from your phone’s hotspot, if your data plan allows it). 

9 - VPN and Security Layers

Some users connect to cloud environments via corporate VPNs or secure gateways. Depending on the setup, this can introduce additional processing steps that may influence performance.

A simple way to test if a VPN is causing issues, is to turn off the VPN and run the tests for latency, bandwidth and for stability and consistency. Do you notice big improvements in bandwidth or a reduction in lag? Then the VPN is the likely culprit. Raise this with your IT department. 

4. Collaboration and File Management: How Cloud Environments Handle Project Data

Revit is often used in collaborative environments, where teams work on shared models stored in BIM 360, Autodesk Docs, or network locations. As you have probably guessed, cloud workstations will access files differently. Spoiler alert: you will likely enjoy big improvements versus your local workstation. 

10 - Shared model access

One of the reasons why cloud-based Revit provides real advantages over workstations is because in the cloud, it is very easy to equip an entire project team with the exact same version of Revit. This prevents Revit crashes and model corruptions, a common source of wasted effort and user frustration. This is much harder to do with on-premise workstations, especially when teams are spread across locations, home office and contractor sites.  

Here’s an extensive review of file collaboration with cloud workstations, which will go much deeper into the 5 different setups that are possible to collaborate with Revit.

We will focus here only on the performance side of things. There are three possible scenarios

A. On-premise File Access

The first possibility is to connect the cloud workstation to a Revit central model file on a local area network. This is, from the cloud workstation’s point of view, the worst option. That means it will probably be slow, just like physical workstations accessing the Revit central model file from outside the network. To connect the cloud workstation safely and securely to the file server where the central model file is located, you’ll need to install a VPN connection that gives the cloud workstation access to the Local Area Network. Not recommended.

B. FileShare Access 

The second option is to have a FileShare (a file server) directly connected in the cloud to the cloud workstation. In this way, the cloud workstations can “talk” directly to each other. This will lead to a performance that is equal to or better than in an office (local area) network.

 Test this setup by opening a Revit model on your cloud workstation from the file server. 

C. Autodesk Construction Cloud/BIM 360

So if your company is using ACC/BIM360, then here's a positive surprise. Both ACC and your Cloud Workstation will be closely located to the internet backbone. Even more important: Cloud workstations have bandwidths that are a factor 10-100 of traditional workstations. Don’t be shocked if you see download speed exceeding 5GB/s. 

To see this in action, do a speed test on your cloud workstation.

Try Revit in the cloud for yourself

If you’re interested in the potential of Revit in the cloud, but are concerned about performance issues, then the best option is simply to try it for yourself. Each company’s Revit needs and setup are unique to them, and so the factors that could cause performance issues for some, might not affect others. With this trial, you can do your own tests and determine how much of an issue performance really is when using Revit in the cloud. 

As mentioned in the introduction,  isn’t the only concern that Revit users have about the cloud. Read our guides on internet access and security for Revit in the cloud, and learn more about how you can address these concerns.

Read the next article on concerns over internet access
Read the next article on Revit cloud security