The network of cloud providers, CDNs, server makers, and software vendors running VPUs in production today. No matter where you run video, a proven deployment path already exists.
No one deploys a chip on its own. A VPU takes over one stage of the video pipeline: encoding. Ingest, packaging and delivery run on other companies’ platforms.
A deployment is only complete when the VPU’s encoded output is connected to the ingest, packaging and delivery systems on either side of it.
19 companies in four groups: CDNs and platforms, servers and systems, video software, and integrators and services.
Every one of them runs VPUs in production or builds products on them. There is no formal program to join. Membership means the integration exists and runs in production. We promote the network because it works.
Every integration on this page is validated and already running, so you are not the first one through. When trouble occurs, the partner will contact us and we troubleshoot together as a team to find the root cause.
NETINT field application engineers often work with the partner and their end customer to root cause issues. We are more concerned with addressing these issues than passing them onto someone else.
Click any partner for details.
Click any partner for details.
Three structural shifts happened in sequence. They are the reason an open-market ecosystem became necessary, not optional.
Software encoding was never designed for today’s codecs or scale. AV1, 4K and 8K expose compute and power limits immediately.
Architectures collapse as scale increases. The math stopped working
The market moved faster than general-purpose compute. Efficiency improvements are incremental. The gap widens every year and compounds.
You cannot optimize your way out of a broken model.
Video infrastructure is now divided between those who own silicon and those who don’t. Walled gardens (Google and Meta) built custom ASICs with 20-30x efficiency advantages.
The open market has zero access to proprietary silicon.
When something goes wrong across a multi-partner VPU deployment, the path is the same every time. Four steps, named owner, system-level resolution.
GPU cracking, the AV1 tipping point, the TCO blind spot, and why hybrid at scale is the dominant emerging architecture.
Nine ecosystem partners. Nine architectural perspectives. The companies building with VPU infrastructure explain why.
The VPU Ecosystem is the set of companies that build, deploy, and integrate VPU-native infrastructure: cloud providers, CDNs, server makers, and video software vendors. Together they cover every architecture, region, and workload, so a deployment path to VPUs already exists wherever you run.
Yes. Ecosystem partner clouds, including Akamai Cloud and NetActuate, offer VPU instances. You encode on purpose-built silicon without owning hardware, and you can move between cloud, owned, and hybrid deployments as your economics change. See our Cloud Pilot Program for details.
Akamai Cloud, Arcadian, CDN77, Cires21, i3D.net, NetActuate and Scalstrm, among others. Every name in the Ecosystem Directory runs VPUs in production or builds products on them. The directory lists who does what.
Start a conversation. There is no formal program and no paperwork wall. Ecosystem companies are the ones that have deployed VPUs in production or built products on them. If you are building on VPU-native infrastructure, you are already most of the way in.
When something goes wrong across a multi-partner VPU deployment, the path is the same every time. Three steps, named owner, system-level resolution.
You file once.
NETINT FAE owns it.
Partner FAEs join the bridge.
Resolved at the architecture.
The Deployment Playbook is how converging companies got there. Get it free or start a conversation with our field team.
The Deployment Playbook is how converging companies got there.
Get it free or start a conversation with our field team.
