Blender Render Farm for Cycles & EEVEE: Rent 1–8x RTX 5090 Servers

• A whole multi-GPU workstation, not a slot in a queue. Rent up to 8x RTX 5090 with 32 GB VRAM each. Every server has 256 GB ECC RAM and a 2 TB NVMe SSD, and the 4x and 8x servers run an AMD Threadripper PRO 5975WX with 32 cores. Cycles renders each frame on all the GPUs you tick in Preferences.
• Your Blender, unchanged. Use the exact Blender build your scene was made with, add your own add-ons and asset libraries, and check the scene in the viewport before the batch. Nothing on the farm side re-reads or converts your .blend file.
• No license per server. Blender and Cycles are free, so you can open one, three or ten servers for the same animation without buying extra render licenses.
• You pay for render time, not for setup. Upload to iRender Drive for free before the server starts. Your first top-up of $50 or more is doubled, which covers setup and test frames. The 2x, 4x and 8x plans bill per minute, and sessions of 3 hours or more get 10% off the whole bill.
• Real numbers before you rent. In our test, 120 Cycles frames at 1080p rendered on 8x RTX 4090 in 45 minutes, about $39 at list price.

Most Blender render farms are submission farms: you upload a .blend file, the farm renders it with its own Blender build and sends frames back. iRender works differently. You rent the whole server and drive it yourself, so your Blender version, add-ons and settings stay exactly as they are.

Heavy Blender scenes usually fail for the same few reasons: the scene does not fit in VRAM, a texture path breaks after packing, an add-on is missing, or the CPU spends minutes on “Synchronizing object” before the GPUs start. The table matches each one to what an iRender server gives you.

The figure gives the short version. The table below goes row by row: how each heavy-scene problem shows up in production, and the exact hardware or access that handles it. If your scene hits more than one row, a 4x or 8x RTX 5090 server (plan 5i or 9i) covers all of them on one machine.

Every plan is a whole server. EEVEE and stills need one GPU; Cycles animations scale with 2, 4 or 8 GPUs. Pick the card for your scene size: RTX 5090 when the scene needs more than 24 GB, RTX 4090 when it fits.

Cycles does not add VRAM across GPUs. The Blender manual is clear that each GPU can only access its own memory, so every card holds the whole scene: the limit is 24 GB on RTX 4090 plans and 32 GB on RTX 5090 plans, however many GPUs the server has. More GPUs make each frame faster; they do not make room for a bigger scene.

When the GPU memory is full, Cycles with CUDA or OptiX automatically uses system memory instead, and leaves at least half of the system RAM to other software. The Blender developers note that this costs performance: in their tests, image textures that did not fit in VRAM rendered about 20 to 30% slower, and when other scene data spills over it can be far slower. Every iRender server has 256 GB ECC RAM for this. If your scene is close to 24 GB, choose an RTX 5090 plan.

These are iRender tests recorded on video, with the server, frame count and render time shown on screen. Costs are worked out at list price for the render time. Each test uses its own scene and settings, so compare a row with your own scene, not with the other rows.

Watch the full 8x RTX 4090 run, from uploading the project to the final frame, including how to enable all 8 GPUs with OptiX and why you should copy the project to the local drive before rendering.

One Blender session rendering on all GPUs is the simplest setup, but light scenes often do not scale well: each frame still waits for CPU work such as scene sync and BVH building, and the GPUs sit partly idle in between. The video below shows how to check GPU load and what to do about it.

On a multi-GPU server you can fix this yourself. Open several Blender instances, tick a single GPU for each in Preferences > System, and give each instance its own frame range. Turn off Overwrite and turn on Placeholders in the Output properties, so no two instances render the same frame. The 2x and 4x RTX 4090 tests above were rendered this way.

Submission farms are popular for animation because they split frames across many nodes. You can do the same on iRender: rent two, three or more servers, open the same project from iRender Drive on each, and give each server its own frame range. Blender needs no render license, so extra servers cost only their own minutes, plus a few minutes of setup on each.

Each server opens the same project from iRender Drive and renders its own frame range, set as Frame Start and End in the Output properties, or with blender -b scene.blend -s 1 -e 40 -a on the command line. Here is what that looks like with the real 120-frame test from the benchmark table.

• Cycles (OptiX or CUDA): uses every GPU you enable, so 2x, 4x and 8x servers cut the time per frame. Choose RTX 5090 plans when the scene is close to 24 GB.
• EEVEE: renders on one GPU, so a 1x RTX 4090 (3S) or 1x RTX 5090 (3i) server is the cost-effective choice. For a long EEVEE animation, rent several 1x servers and split the frame range.
• Other engines in Blender: with full administrator access you can also install Octane or Redshift for Blender with your own license. See the Octane render farm and the Redshift render farm.

1. Register and top up. Your first top-up of $50 or more is doubled, so setup time is covered.
2. Upload your project to iRender Drive. Transfer and storage are free, and you can upload before the server starts.
3. Start a server and connect over remote desktop. Install your Blender version and add-ons, copy the project to the local drive, and tick the GPUs in Preferences > System.
4. Render with F12, Render Animation or the command line, download the frames and shut the server down; billing stops.

Connect to your server from whatever machine you work on. The iRender GPU app runs on Windows, macOS (Apple Silicon and Intel) and Ubuntu/Debian, and the web app opens in any modern browser with nothing to install, so a Mac user can drive an 8x RTX 5090 Windows server.

Once connected, the server behaves like a local workstation. The quickstart below shows the full flow on iRender, from uploading the project to shutting the server down.

A multi-hour 8x GPU render needs the building around it to stay up, and studio work under NDA needs a provider that can show its security controls. iRender servers run in an Uptime Institute Tier III data center with ISO-certified operations.

Support is available 24/7 on weekdays and from 6:00 AM to 11:59 PM (UTC+7) on weekends, by hotline, Skype and email.

Hand-picked guides from the iRender blog for the problems Blender artists ask us about most: preparing scenes, missing textures, GPU backends and slow scene sync.