NETINT · Creator of the VPU

Video encoding silicon.
Built for scale.

A VPU is a chip that does one job: encode video. NETINT invented it. Run VPUs in servers you own or on partner clouds using open source and commercial software integrations.

225,000+

VPUs in production

1.1 Trillion+

Minutes of video encoded

Close-up of a black textured hardware enclosure with NETINT and QUADRATI labels and red cables.

NETINT · Creator of the VPU

Video encoding silicon.
Built for scale.

A VPU is a chip that does one job: encode video. NETINT invented it. Run VPUs in servers you own or on partner clouds using open source and commercial software integrations.

225,000+

VPUs in production

1.1 Trillion+

Minutes of video encoded

Close-up of a black textured hardware enclosure with NETINT and QUADRATI labels and red cables.

Why a VPU?

A VPU spends its whole chip on one job: encoding video. CPUs and GPUs spend most of their silicon on other work.

How much of each chip encodes video

Least optimized

CPU

Video silicon: none

General-purpose cores encode video in software.

Partly optimized

GPU

Video silicon: a small block

Built for graphics and AI. Encodes using a small block of the chip.

Fully optimized

VPU

Video silicon: the whole chip

Every block on the chip encodes and decodes video.

Cost per stream climbs on CPUs and GPUs.
On VPUs it stays flat.

Three changes pushed video encoding off general-purpose chips.

1

Codecs got heavier

AV1 and HEVC make smaller files, but each frame takes far more computing to encode. In software, that burdens CPU cores. VPUs encode AV1 in fixed-function logic for the same costs as encoding HEVC or H.264.

2

Stream counts multiplied

Platforms now encode thousands of streams at once due to ladder and codec variations. On CPUs, cost climbs with every added stream. One 1RU Quadra server can encode 320 streams.

3

Power became the limit

A data-center rack runs out of power before it runs out of slots. A VPU card encodes 32 HD streams at 17 W. A 1RU Quadra server encodes 320 streams on less than 500 W.

MEASURED, NOT MODELLED

Tested against CPU and GPU.

VPU compared to x265

7.6xgreater density

Quadra
19.0
NVENC P7
8.5
x265 medium
2.5

HEVC at 32 concurrent 1080p sessions on identical Akamai hosts.

Source: NETINT Akamai Throughput and Quality Report, SDK V5.7.2.

VPU compared to GPU

4.7xfewer watts

Quadra T1U
13 W
RTX 4000 Ada
51 - 82 W

Peak VPU power 13 W against 51 to 82 W for an NVIDIA RTX 4000 Ada on a live ABR ladder.

Akamai independent test results, Frankfurt, 19 August 2025.

VPU compared to x264

3xenergy efficiency

VPU
2 - 5 W
CPU
15 - 20 W

15 clips, H.264, HEVC and AV1, energy logged through OpenBMC.

Independent test by G&L Systemhaus, 9 July 2025.

PRICE AND PERFORMANCE

Cost per live HD stream, per month.

Same 1080p channel. Three ways to run it.

CPU, 8-core instance
$72.00
AWS VT1 (Xilinx U30)
$59.38
VPU (2× Quadra T1U, Akamai Cloud)
$27.11

Monthly instance list price divided by sustained live 1080p streams: 2 on the CPU plan, 8 on VT1, 18 on the VPU plan. September 2025 prices. Your inputs will differ.

Where VPUs run

Built for the video you run.

Thousands of concurrent encodes, AV1 and HEVC, at a flat cost per stream.

Peak-event capacity with low-delay encoding and no rebuild of the pipeline.

Encoder latency under 5 ms at 1080p for gaming, betting and auctions.

SDI in, IP out, in a 1RU server or a half-rack appliance.

Decode, downscale and run inference on hundreds of camera streams per chip.

VPU Products: purpose-built silicon.

SILICON

M.2, U.2 and PCIe form factors. 20 to 64 HD streams per card for servers you own.

Red Quadra video processing unit by Netint with a PCIe edge connector on the left, shown with a black studded NetInt accelerator box on the right.

SERVERS

1RU with ten VPUs, 320 HD or 80 4K streams. Half-rack Mini Server for the edge.

Rack-mounted storage server with eight hot-swappable drive bays.

Software

Bitstreams

The control software that monitors VPU silicon performance and stream management.

Laptop displaying a Bitstreams dashboard with left navigation and multiple data panels for activity and hardware monitoring.

Cloud

Cloud VPUs

The same card on Akamai Cloud, i3D.net and NetActuate VPUaaS.

Three rack-mounted server enclosures in a data center, filled with hot-swappable drive bays and indicator lights to the front.

Size your deployment

Send us your stream count, codecs, resolutions and latency target.

A NETINT engineer will spec the cards, servers or cloud instances
to match, and show the power and rack-space math.

Send us your stream count, codecs, resolutions and latency target.

A NETINT engineer will spec the cards, servers or cloud instances to match, and show the power and rack-space math.

LEADERSHIP AND INNOVATION

A decade building the VPU category.

2015

NETINT founded in Vancouver, Canada to build video-only silicon

2019

Codensity G4, Logan ships as the T408 and T432

2021

Codensity G5, the Quadra family adds AV1, 8K and on-chip AI engine

2024

Technology and Engineering Emmy Award, alongside Google, Meta and AMD

2025

Akamai becomes the first infrastructure for video cloud to offer VPUs

2026

Over 225,000 VPUs in production and 1.1 trillion minutes of video encoded

Decorative gold ribbon sculpture with intertwined loops on a black background

VPU Emmy®, alongside Google, Meta and AMD.

The 2024 Technology and Engineering Emmy Award recognized NETINT for Design and Deployment of Efficient Hardware Video Accelerators for Cloud.

What only a video-first chip can do

Encode AV1 at the same density as HEVC and H.264.

Decode, downscale and run an AI model on the same die.

Move frames straight from a GPU into the encoder.

Fit ten cards in one rack unit, enabling 320 live channels at broadcast quality.

VPU Ecosystem Partners

CDN & Platform

Software

Integrators & Services

Servers & Systems

Choose your entry point.

Explore →

1 to 2 weeks

Test →

2 to 4 weeks

Migrate →

4 to 8 weeks

Scale

ongoing

We de-risk VPU migration with a proven four-stage process that's parallel and reversible.

Search latest news and info.

Buyer's guide

How to Specify a VPU: Buyer's Guide

Form factors, measured latency and throughput, VMAF curves, a requirement-to-product map.

Customer story

Zapping: 100 servers to 25

A streamer’s production fleet was cut 75% using VPUs to replace CPU and GPU-based encoding workflows.

Blog

Stop Asking Which Encoder.

A five-step decision sequence with Akamai performance data.

Technical walkthrough

The FFmpeg walkthrough, command by command.

Industry report cover for 2026: The State of Video Encoding with subtitle about codec adoption and streaming infrastructure; NETINT logo bottom-right.

Industry report

The State of Video Encoding 2026

Netint hero banner with the headline “The VPU Deployment Playbook” over a red gradient skyline and blue light-trail road imagery.

TECHNICAL GUIDE

The VPU Deployment Playbook

WHERE TO FIND NETINT

Upcoming events.

APAC

Inter BEE

November 18–20

WHY MEET NETINT

Meet NETINT at Inter BEE to power IP, cloud and AI media workflows in Japan with high-density, energy-efficient video encoding and transcoding.

NORTH AMERICA

Sports Video Group (SVG) Summit

December 15–16

WHY MEET NETINT

Meet NETINT at the SVG Summit to power live sports production and distribution with low-latency, high-density video encoding at lower cost and power.

NORTH AMERICA

CES (Consumer Electronics Show)

January 6–9

WHY MEET NETINT

Meet NETINT at CES to see how VPUs power efficient video encoding for next-generation streaming devices and consumer media, more channels, less power, lower cost.

Start encoding on a VPU this week.

In the cloud

Start your VPU proof of concept to gather real world data around your workflow.

In your own server

See how VPUs can accelerate your existing and decommissioned servers.

Talk to a VPU engineer

Schedule a short call to match your workload to a VPU and a test plan.

TAKE OUR SURVEY

State of Video Encoding 2027

Tell us where you see the industry going next? Submit for a chance to win a Quadra T1U VPU.