AV1 is the practical production choice in 2026. AV2 may ultimately deliver better compression and more flexible media experiences, but its newly finalized specification is only the starting point for optimized software, hardware, browsers, devices, packaging, and commercial deployment.
For most streaming teams, the decision is not AV1 or AV2 on equal terms. Use AV1 where its efficiency justifies the encoding and device-compatibility work. Treat AV2 as an evaluation target for codec development, hardware planning, and future services until its end-to-end delivery ecosystem matures.
AV1 vs AV2: The Short Answer
- Choose AV1 when you need a royalty-free, efficiency-focused codec that can be deployed across an expanding ecosystem of software, browsers, devices, and streaming workflows.
- Evaluate AV2 when you are planning for higher compression efficiency, multi-stream or multi-view experiences, screen content, AR or VR, or future codec deployments.
- Monitor AV2 without adopting it prematurely when you want to understand its future potential but your service depends on mature encoders, decoders, hardware, players, and devices today.
What Are AV1 and AV2?
AV1, or AOMedia Video 1, is an open video codec released in 2018 by the Alliance for Open Media. It was designed for efficient internet video delivery under AOMedia’s royalty-free patent policy. AV1 now has production encoders and decoders, browser and operating-system support, hardware acceleration on many newer devices, and documented deployments by major streaming and communication services.
AV2, or AOMedia Video 2, is AV1’s successor, but it defines a new bitstream and decoding process rather than a backward-compatible AV1 profile or decoder upgrade. Version 1.0.0 was finalized on May 28, 2026, and publicly announced on June 9, 2026. It targets higher compression efficiency, scalable bitstreams, multi-stream and multi-view video, screen content, broadcasting, streaming, and real-time video conferencing.
A final codec specification is not the same as a production-ready delivery ecosystem. The final AV2 release gives implementers a stable technical target, and the AOMedia Video Model reference software is available. Optimized encoders and decoders, container bindings, hardware acceleration, browsers, operating systems, players, and shipping devices must still mature around it.
Who Already Uses or Supports AV1?
AV1 has been used in production for years. Major streaming platforms, communication services, device manufacturers, and software ecosystems have deployed or documented AV1 support. The AOMedia AV1 adoption showcase includes implementations from Google, Netflix, Meta, Microsoft, Vimeo, VideoLAN, and other organizations. These examples do not mean that every viewer receives AV1, but they demonstrate a real production ecosystem that AV2 does not yet match.
- Netflix: Netflix reported in December 2025 that AV1 powered approximately 30% of all viewing on its service. Netflix also described AV1 HDR streaming and AV1 delivery across Android, smart TVs, browsers, and other large-screen devices. This is evidence of large-scale production deployment rather than a limited test. Read Netflix’s AV1 deployment update.
- Meta: Meta has published results from AV1 deployment for Facebook and Instagram Reels, including a 12% average bitrate reduction and improved playback quality in one rollout. In 2026, Meta also reported at least a 20% bitrate reduction compared with H.264/AVC in offline real-time communication tests on low-end and mid-range devices. The results show why AV1 can be valuable for both stored video and real-time communication, while also highlighting the need to manage encoding complexity. See Meta’s Reels case study and Meta’s RTC update.
- YouTube and Google: AOMedia documents AV1 use across YouTube, Chrome, Android, and Google Meet. This directly contradicts any suggestion that AV1 is still primarily a laboratory technology. It is part of a broad production platform ecosystem, although actual delivery still depends on content, client capability, and product policy. Review AOMedia’s Google adoption story.
- Android and hardware manufacturers: Android added AV1 platform support in Android 10 and continues to improve software decoding for devices without dedicated AV1 hardware. Intel documents AV1 hardware decode support across newer processor generations, while Samsung introduced built-in AV1 playback support in its 2020 8K QLED TV lineup. Hardware coverage remains device-specific, but these examples demonstrate that AV1 playback is already available across important parts of the consumer ecosystem. Read Android’s AV1 support announcement, check Intel’s codec capabilities, and see Samsung’s AV1 announcement.
These deployments also explain why AV1 remains the practical codec for production decisions in 2026. Teams can test it against real devices, players, networks, and operational metrics. AV2 may ultimately improve compression and support more advanced media workflows, but it must first develop a comparable implementation and hardware ecosystem.
AV1 vs AV2: Side-by-Side Comparison
Compression Efficiency and Video Quality
Compression efficiency describes how much visual quality a codec can preserve at a given bitrate, or how much the bitrate can be reduced while maintaining a similar quality level. A published evaluation of the pre-final AV2 v13.0 development codec, using AV2 Common Test Conditions, reported a 29.81% BD-rate reduction with PSNR-YUV and a 33.79% reduction with VMAF under its random-access configuration, compared with its AV1 test anchor. These results are an important indicator of AV2’s potential, but they are not a benchmark of final AV2 v1.0.0 implementations. Review the AV2 evaluation methodology and results.
Those figures are controlled benchmark results, not a guaranteed saving for every source, product, or live workflow. Actual performance depends on content, resolution, frame rate, color depth, encoder implementation, preset, latency constraints, rate-control mode, and quality metric. The study also evaluated development software rather than mature commercial AV2 encoders. Teams should test representative footage with objective metrics and human visual review.
The practical promise is straightforward: AV2 may let a provider deliver the same quality with fewer bits, or improve quality without increasing the bitrate. That could reduce bandwidth and storage costs for large video libraries. It could also make high-resolution, multi-view, and immersive video more practical. The benefit will depend on whether the encoding and playback ecosystem can support the format efficiently.
Encoding, Decoding, and Hardware Support
AV1 adoption has been slowed by the cost of encoding, the need for hardware decoding, and uneven support across devices. Dan Rayburn’s summary of NETINT’s 2026 State of Video Encoding survey identifies hardware decode support, toolchain limitations, and encoding compute costs as the leading barriers to AV1 adoption. Those same operational questions matter even more when evaluating a newer codec.
AV2 now has a stable specification and reference software, which gives encoder and decoder developers a foundation for implementation. AOMedia says the next phase includes optimized software, broader hardware support, conformance testing, container bindings, and commercial product integration. The AOMedia implementation update explicitly says software implementations typically come first and that hardware support, which is essential for broad device adoption, follows.
Independent AV2 decoder development has begun. VideoLAN’s May 31, 2026 announcement described its first dav2d release as feature-complete for AVM v15, with 8-bit and 10-bit decoding and initial optimization work for x86, Arm, and RISC-V. VideoLAN also described the project as early work focused on correctness, conformance, optimization, and platform support. This is evidence of implementation progress, not final AV2 v1.0.0 conformance or broad production readiness.
AV2 Hardware and Ecosystem Signals in 2026
Early AV2 hardware activity is real, but the distinction between licensable decoder intellectual property and a shipping consumer device matters. Semiconductor IP announcements show that system-on-chip development can move forward. They do not prove that AV2 decoding is already available across production GPUs, phones, televisions, browsers, or streaming devices.
- VeriSilicon: The company announced on June 9, 2026 that its VC9800D video processing unit IP supports AV2 decoding and had been provided to multiple customers.
- Chips&Media: The company announced on June 10, 2026 completed AV2 hardware decoder IP and later disclosed a licensing agreement with an unnamed North American technology company. No shipping consumer product was identified.
- Allegro DVT: The company announced real-time AV2 decoder IP in its Pulsar D400 series on June 16, 2026.
AV2 will also receive a formal industry update at IBC2026. The Coding the Future session, scheduled for September 12, 2026, includes an AOMedia update on AV1 and the newly published AV2 codec, followed by discussion of market impact and next steps. Andrey Norkin, Principal Architect at NVIDIA and co-chair of AOMedia’s Video Codec Working Group, is a confirmed speaker.
Features and Use Cases
AV1 is a strong fit for internet video delivery where bitrate efficiency, storage, and royalty-free codec access matter. Common use cases include video on demand, high-resolution streaming, mobile delivery, user-generated content, and adaptive bitrate ladders that serve AV1 to compatible clients while using a fallback codec for the rest of the audience.
AV2 keeps those goals and expands the design for more complex media. The official specification and AOMedia’s release materials describe support for scalable bitstreams, multiple synchronized streams, multiple camera angles, stereoscopic video, composite video, screen content, AR, VR, broadcast, and video conferencing. These capabilities can reduce the application-level coordination required to deliver related video components together.
AV2 may therefore be especially interesting for immersive experiences, interactive sports, multi-camera production, remote collaboration, and services that need to combine video components. The technology does not remove the need for application design, packaging, player support, or device testing. It provides a more flexible codec foundation for those workflows.
What Video Teams Are Using in 2026
Dan Rayburn’s report on NETINT’s 2026 State of Video Encoding survey shows why AV1 is worth evaluating before AV2 becomes widely available. Rayburn writes that AV1 has 17% current production deployment, while 40% of respondents plan to deploy it in 2026. Together, those figures represent a 57% combined reach by the end of the year, although they are survey results rather than a measurement of total market share. Rayburn describes this as mainstream planning rather than early-adopter experimentation.
NETINT sponsored the survey and distributed it through industry channels and its own network. Its sample may overweight organizations already evaluating hardware encoding, and APAC and LATAM are underrepresented. Treat the figures as directional evidence from the surveyed organizations, not a representative measure of the whole video market.
The survey also shows that codec adoption is connected to operational maturity. Organizations running three or more codecs in production are reported to be 57 times more likely to have AV1 in their stack than single-codec operators. Teams that already manage multiple encoders, device profiles, testing workflows, and delivery ladders are better positioned to add another codec.
The closest published survey covering AV2 comes from AOMedia itself. In September 2025, AOMedia reported that 53% of surveyed members planned to adopt AV2 within 12 months of finalization and 88% expected to implement it within two years. This measures roadmap intent among AOMedia members, not independent market share or confirmed product shipments, so it should be treated as an ecosystem signal rather than an adoption forecast.
The reported barriers to AV1 adoption are primarily operational. Hardware decode support leads at 54%, followed by toolchain limitations at 43% and encoding compute costs at 37%. Those findings reinforce the central lesson for AV2: compression efficiency alone does not determine deployment. Hardware coverage, tool maturity, encoding economics, packaging, playback support, and operational capacity determine whether a codec is viable at scale.
Which Codec Should You Use?
- Use AV1 now if you need a more efficient delivery codec, your target clients support AV1, and your encoding infrastructure can meet the required throughput and latency.
- Evaluate AV2 experimentally if you are researching immersive or multi-view experiences, developing codec technology, or want to understand its future capabilities. Evaluation is different from production adoption.
- Keep a fallback if your audience includes older devices or your service requires predictable playback across many platforms. Codec ladders are usually more practical than a single-codec rollout.
- Do not choose AV2 only from a compression claim. Compare the total system cost, including encoding, decoding, hardware, tooling, packaging, monitoring, playback, and support.
AV1 codec is now supported for ingest and egress streams across all protocols supported by Red5 Pro. AV2 support should be confirmed separately with each server, encoder, protocol, player, and device vendor because the codec’s first-generation ecosystem is still developing.
How to Evaluate AV2
Evaluating AV2 does not require an immediate migration or production rollout. Start by documenting your current codec stack, encoding presets, bitrate ladders, packaging formats, delivery protocols, player capabilities, device distribution, and quality metrics. This baseline will make future AV2 tests comparable and will show where the new codec could create value.
Next, create a test plan using representative live and on-demand content. Measure compression efficiency, encoding throughput, decoder performance, startup time, stream switching, latency, power use, playback stability, and infrastructure cost. Include the devices and network conditions that matter to your audience. Compare AV2 against your actual AV1 configuration, not only against a slow reference preset.
Before production adoption, require evidence of an optimized encoder and decoder, stable packaging and protocol support, browser or native-player integration, hardware coverage across target devices, conformance testing, monitoring, and a tested fallback. A reference implementation or hardware IP announcement can demonstrate progress, but neither proves that an end-to-end service is ready to replace AV1.
FAQs
Is AV2 better than AV1?
AV2 is designed to be more compression-efficient than AV1 and to support advanced multi-stream, multi-view, screen-content, AR, and VR workflows. However, AV1 has a more mature production ecosystem, broader hardware availability, and more proven deployment options. AV2 is technically newer, but it is not automatically the better operational choice for every service today.
What do AV1 and AV2 mean?
AV1 means AOMedia Video 1, an open video codec developed by the Alliance for Open Media. AV2 means AOMedia Video 2, its successor. AV2 defines a distinct bitstream and decoding process rather than a backward-compatible AV1 profile, while targeting higher compression efficiency and more flexible multi-stream, screen-content, broadcasting, and real-time workflows.
Is AV2 royalty-free?
AV2 was developed under AOMedia’s royalty-free patent policy, and AOMedia describes the specification as a royalty-free video coding standard. Organizations should still review the current patent policy, implementation terms, and legal requirements for their specific product. Royalty-free codec design does not remove the need for normal legal and technical due diligence.
Why is AV1 not more popular?
AV1 adoption has been slowed by hardware decoding coverage, toolchain limitations, encoding compute requirements, and uneven support across older devices. Its production use is growing, but deployment requires testing and operational work. AV1 is most practical when a team can control the delivery ladder, target compatible clients, or justify the efficiency gains at scale.
Conclusion
AV2 is a meaningful next step in open video coding. It is designed to deliver better compression efficiency than AV1 and to support more flexible multi-stream and immersive video workflows. The specification is now available, but the practical value of AV2 will depend on encoders, decoders, hardware, players, devices, protocols, and production tooling.
AV1 is the safer production choice for most teams in 2026. Use it where its efficiency and royalty-free design solve a real delivery problem, and measure the results across your actual devices, networks, and workflows. AV2 is worth tracking and evaluating, but its first-generation ecosystem is still far from broad production adoption.
Top comments (0)