Dedicated GPU machines that sit in the same data centres as your delivery nodes and run your encoding pipeline. We supply and operate the hardware. What runs on it stays yours.
Reply within 3 business hours. No sales sequence.
Worth saying plainly, because the two are often sold as one thing and they are not.
The machine and everything under it: GPUs, CPU, memory, NVMe, transit and the data centre. We rack it, patch it, monitor it and replace it when hardware fails. On a self-managed build you get root and a node that behaves like the rest of your infrastructure. A managed build is also possible, and there the access model is written into the agreement rather than assumed.
We do not transcode your video for you, and we do not sell encoding as a service. Your ladder, your codecs, your packaging rules and your quality decisions stay in your pipeline, where they belong. Our delivery product says the same thing: Video CDN stores and serves your files exactly as you upload them.
This is custom infrastructure ordered per build. It is not a feature of Video CDN and it is not bundled with delivery.
A mezzanine file is large. Every time it crosses a vendor boundary, somebody bills for the crossing and the clock keeps running.
Encode where the output already needs to live. The rendition does not have to travel from a compute vendor to a delivery vendor before anyone can watch it.
When the encoder finishes inside the same network that will serve the file, the gap between the last frame written and the first frame delivered is a short one.
When a release is late, the compute and the delivery are on one contract with one support channel, so nobody spends the morning deciding whose incident it is.
Whatever you already use. A self-managed node with root access does not care which encoder you picked, and neither do we.
FFmpeg and the rest of the usual chain, built the way your team builds it, with the GPU flags your ladder needs.
If you have licensed software and an operations team that knows it, the node is somewhere to run it rather than a reason to change it.
Teams that have spent years tuning a pipeline rarely want a vendor’s opinion about it. Bring the container, keep the tuning.
The same class of hardware handles rendering and inference. If that is the larger half of your workload, start from GPU Servers for AI Projects instead.
NVIDIA RTX PRO 6000 Blackwell Server Edition is the card most video pipelines land on, because it handles encoding and the graphics or inference work sitting next to it on the same node.
A working ladder for a catalogue that grows steadily, or a first node beside an existing pipeline.
For a back catalogue being re-encoded in bulk, or live ladders running alongside on-demand work.
Larger multi-GPU configurations exist and are specified per build. GPU pricing moves with the AI market month to month, so we quote the current rate rather than print one that goes stale. The full line sits on Custom Enterprise CDN Infrastructure.
The renditions your node produces have to live somewhere and then reach viewers. Both of those steps are priced openly, so you can add them up before you talk to us.
Your pipeline finishes a rendition on local NVMe. Nothing has left the building yet, and nothing has been billed for movement.
Cloud Storage runs at $0.015 per GB. If the file should stay resident in the network instead, permanent cache is $0.02 per GB with the first 500 GB included.
Delivery runs on the standard tiers, $5.00 down to $2.50 per TB. A month of 100 TB comes to $415. See pricing.
Only the node itself is quoted per build. Everything downstream of it is on the published price list, which is the part most compute vendors leave for you to discover on the invoice.
Two different products that people ask for with the same sentence. They solve opposite halves of the job.
Your problem is making the renditions. Encoding is slow, or it costs too much, or the output has too far to travel before anyone can watch it. Custom infrastructure, quoted per build.
Your renditions already exist and your problem is serving them. Files replicate inside the network and are delivered on the standard tiers, untouched. See Video CDN.
Plenty of teams end up with both, and then the two sit in the same data centres by design. Related: Infrastructure for Streaming.
The six questions that come up on every scoping call about processing hardware.
On a self-managed build, you do. We hand over a node with root access and a working network. The software, the licences and the pipeline are yours, which is also why we do not charge you for them.
That depends on your ladder, your codec and your preset, and any number quoted without those three is marketing. Bring your ladder to the call and we size against it.
A fixed monthly figure, quoted per build, following the GPU count, the CPU, the memory and the term. GPU pricing moves with the AI market, so we quote the current rate instead of publishing a stale one.
The processing nodes sit in the same data centres as delivery nodes, so the output does not cross a vendor boundary to get into the network. Figures for your specific locations come with the quote.
Set per build, because it follows the hardware and the data centre commitment behind it. Whatever it turns out to be, it is in the quote before anything is ordered.
Often, yes, and we will say so. The case for a node here is that the output lands next to the network that serves it. If your renditions are small or infrequent, that adjacency is not worth paying for.
Tell us what your pipeline does and how much of it there is. You get a reply within 3 business hours, from an engineer who will say if a general cloud instance would serve you just as well.
He reads every request sent from this page and replies within 3 business hours, including the reply that says a general cloud instance would serve you just as well. No sales sequence in between.