Licensing & Economics

The Problem: A Thicket of Patents

Modern video codecs are complex — so complex that no single company holds all the essential patents. A single video standard may implicate thousands of essential patents from hundreds of patent holders spread across dozens of countries.

The core problem is royalty stacking: if twenty different companies each demand $0.10 per device, the total royalty exceeds the profit margin on the device itself. Without coordinated licensing, the transaction costs alone — negotiating individual licenses with hundreds of patent holders — would be crippling.

MPEG LA & the Patent Pool Model

The patent pool was the solution. Instead of negotiating with hundreds of patent holders individually, a neutral administrator bundles patents into a single license. The most famous administrator is MPEG LA (now Via Licensing Alliance), which pioneered the modern video codec patent pool.

The model is brilliant in theory: one-stop shopping, transparent pricing, reduced litigation. In practice, it works well when there's a single dominant pool. When multiple pools compete for the same standard, the model breaks down — a lesson H.265 would teach the industry the hard way.

H.264 / AVC — The Gold Standard

The H.264 patent pool managed by MPEG LA is arguably the most successful technology licensing program in history. It established a predictable, affordable model that enabled the entire streaming video industry.

The annual cap was crucial. It meant that after a certain volume, the marginal cost of adding H.264 support became zero. This gave large manufacturers like Apple, Samsung, and Sony the confidence to build H.264 into billions of devices. The caps also made H.264 essentially free for content distributors — Netflix, YouTube, and Amazon could stream H.264 without per-stream royalty obligations.

H.265 / HEVC — The Licensing Disaster

If H.264 licensing was a well-designed highway, H.265 licensing was a demolition derby. The standard promised 50% better compression than H.264, but the licensing uncertainty nearly killed its adoption.

Three Competing Pools

A device manufacturer implementing HEVC potentially had to pay all three pools — and it was unclear whether the pools covered all essential patents. Some patent holders (like Nokia and Siemens) held out entirely, demanding separate licenses.

The result: adoption paralysis. Apple added HEVC decode in 2017 (iOS 11, macOS High Sierra) but only for their own ecosystem. Android device makers delayed HEVC hardware for years. Netflix and YouTube used HEVC selectively. The licensing mess was arguably the single biggest factor driving the industry toward royalty-free alternatives.

VP8 & VP9 — Google's Royalty-Free Gambit

In 2010, Google acquired On2 Technologies and with it the VP8 codec. They released VP8 as royalty-free — an explicit challenge to the MPEG LA licensing model. VP8's compression wasn't competitive with H.264, but the principle was established.

VP9 followed in 2013, and this was the game-changer. VP9 offered compression approaching HEVC's efficiency — completely free of patent royalty obligations. Google's strategy had three components:

• Defensive patent portfolio: Google amassed a huge patent portfolio and pledged not to assert it against VP9 implementations. Anyone who attacked VP9 users would face Google's patents.
• YouTube integration: Google converted YouTube to VP9, instantly making it the most-watched video format on the planet. Browser vendors had to support it or break YouTube.
• Browser bundling: Chrome shipped VP9 support natively. Edge (Chromium-based) followed. Firefox added support. Safari remained H.264/HEVC-only for years.

The catch: VP9 had no hardware decode support in most devices for years. Apple never added VP9 hardware decoding to their chips (they bet on HEVC). This meant VP9 playback on iPhones/iPads required software decoding — draining battery and limiting adoption on those platforms.

AV1 — The Alliance for Open Media

The Alliance for Open Media (AOMedia) was founded in 2015 by Amazon, Cisco, Google, Intel, Microsoft, Mozilla, and Netflix. Their mission: build a next-generation, royalty-free video codec that would match or exceed HEVC's compression without the licensing nightmare.

AV1 was released in 2018. Its patent model is unique:

• Patent grants: All AOMedia founding members grant a reciprocal, royalty-free license for patents essential to AV1. If you sue someone over AV1 patents, your own patent grant is revoked.
• Defensive suspension: If a member attacks another member over AV1 patents, the attacker loses access to the defensive patent pool.
• Open to all: Any implementer gets a royalty-free license, even non-members — as long as they don't assert AV1-essential patents.
• Third-party patents: AOMedia can't license patents from non-members. If a patent holder outside AOMedia claims essential patents, they can demand royalties — but they'd face the combined defensive portfolio of all AOMedia members.

AV1's compression efficiency is impressive — about 30% better than HEVC and 50% better than H.264. But it came at a cost: AV1 encoding is computationally expensive, requiring 10–100× more CPU time than HEVC encoding. Hardware decoding support is growing but still trails HEVC.

AV2 — Continuing the Royalty-Free Tradition

The AOMedia is already developing AV2, the successor to AV1, with a target of 30% additional compression gains over AV1. AV2 continues the same royalty-free licensing model established by AV1.

The AV2 project incorporates ideas from both traditional block-based coding and emerging neural network-based approaches. It's expected to be finalized around 2026–2027, with hardware decode becoming mainstream in the late 2020s.

Cisco's OpenH264 — A Clever Licensing Workaround

In 2013, Cisco did something remarkable: they published the H.264 encoder/decoder source code as open source (BSD-licensed) and agreed to pay the MPEG LA royalties on behalf of anyone who used their binary module. Mozilla adopted Cisco's OpenH264 binary for Firefox, solving the H.264 support problem without Firefox having to license the patents.

This was a clever legal hack that solved a real problem: Mozilla couldn't afford H.264 licensing, but Firefox needed H.264 support to play web video. Cisco's solution — pay the pool on behalf of the entire open-source ecosystem — was a stopgap that demonstrated both the problem with the licensing model and a potential path forward.

OpenH264 only covers the baseline profile of H.264, with limited compression options. It's not a full codec — it's a legal bridge that allowed open-source browsers to keep up while the industry transitioned to royalty-free alternatives.

Impact on Browsers — The Codec Wars

Browser support for video codecs has been heavily influenced by licensing economics:

The pattern is clear: browsers aligned with companies that had favorable licensing positions. Google (Chrome) pushed VP9 and AV1 aggressively. Mozilla (Firefox) used the OpenH264 hack and embraced VP9/AV1. Apple (Safari) stuck with the licensed MPEG codecs their hardware supported and was slow to adopt royalty-free options. Microsoft (Edge) eventually followed Google after switching to Chromium.

CDN Cost Implications — The Billion-Dollar Savings

Every bit of compression improvement directly reduces Content Delivery Network (CDN) costs. For the largest streaming platforms, even a 1% bitrate reduction saves millions of dollars annually.

For Netflix, moving from H.264 to HEVC saved roughly 35% on bandwidth. Moving to AV1 saves an additional ~15–20% over HEVC. With Netflix spending an estimated $1–2 billion annually on CDN bandwidth, these percentage improvements translate to hundreds of millions in savings. This is why Netflix was a founding member of AOMedia and aggressively pushed AV1 adoption — despite AV1's encoding cost, the CDN savings dwarf the compute cost.

Content Creation Licensing

Licensing isn't just for playback devices — it also affects content production. Professional video production tools that encode video must license the codecs they use:

• Encoding software: FFmpeg, x264, x265 — commercial users need licensing. FFmpeg itself is LGPL/GPL, but patent licenses are separate from copyright licenses.
• Production hardware: Broadcast encoders, studio-grade video gear — manufacturers pay per-unit royalties.
• Content licensing: Distributing content in a licensed format may incur content-side royalties (e.g., H.264 content with >100K subscribers). In practice, most streaming services and broadcasters qualify for the free tier or pay modest fees.
• Royalty-free production: AV1 and VP9 eliminate content licensing entirely — you can encode and distribute without any royalty reporting or payment.

Codec Licensing Cost Comparison

The gap in this table tells the whole story. Royalty-free codecs (VP9, AV1) eliminate licensing friction entirely. The traditional licensed codecs (H.264, H.265) impose per-device taxes that shape hardware and software decisions. The next generation (VVC, EVC) remains uncertain — and the industry is voting with its feet.

Patent Litigation Landscape

Video codecs have generated some of the most expensive patent litigation in technology history. Here are the landmark cases:

The litigation landscape is a major reason why the defensive patent aggregation model of AOMedia is so attractive. By pooling patent commitments from the largest tech companies, AOMedia creates a mutual assured destruction scenario: anyone who sues an AV1 implementer faces counterclaims from the combined patent arsenal of Google, Microsoft, Amazon, Apple, Netflix, and others.

Per-Unit Royalties and Hardware Decisions

Per-unit royalties have a profound effect on hardware design decisions, especially for low-cost devices:

• Streaming sticks ($25–$50): A $0.20 H.264 royalty is manageable (0.4–0.8%). A $1.50 HEVC royalty is 3–6%. At these margins, HEVC support becomes a significant cost factor.
• Set-top boxes ($50–$200): The same logic applies — adding HEVC adds $1–2 in royalty cost, which may push a product out of its target price point.
• Mobile SoCs: Qualcomm, MediaTek, and Apple include HEVC hardware decode in their SoCs, but the royalty cost is rolled into the chip price. For the SoC vendor, paying MPEG LA means they can advertise "HEVC decode" as a feature.
• TVs ($200–$2000): TV manufacturers pay per-unit royalties for all codecs. A 2020-era smart TV might pay $0.20 for H.264 + $1.20 for HEVC + additional for other codecs — adding up to real money across millions of units.

The Bottom Line

Video codec licensing is not a technical decision — it's a business decision that shapes technology adoption more than the technical merits of the codec itself.

The industry is converging on a clear direction: royalty-free for distribution, licensed for production tools. AOMedia's AV1 and AV2 will dominate streaming and communication. VVC/H.266 may survive in broadcast and professional production where per-unit costs are less relevant. The patent pool model that enabled the streaming revolution is slowly being replaced by an open, collaborative alternative — driven not by technical superiority, but by the simple economics of removing friction from the video ecosystem.