AV1 in Production: When Royalty-Free Codec Saves Money (2026)

Why Fora Soft wrote this playbook

Fora Soft has shipped 50+ video streaming and broadcasting projects since 2005. StreamLayer, WorldCastLive, Mangomolo, Perspire TV, Watchfun, the Australia Sports Broadcasting App, plus several large IPTV/STB platforms.

Through 2024–2025 we ran AV1 production migrations for two OTT clients and AV1 evaluations for four more. The numbers in this guide are from those migrations — encoding cost, CDN savings, viewer device coverage, break-even points — not vendor whitepapers.

If you are evaluating AV1 for VOD or live streaming, this guide tells you when it pays off, when it does not, what the encoder pipeline looks like, and how to ship a multi-codec ladder without breaking older viewers.

TL;DR — when AV1 saves money in 2026

AV1 saves money when the same encoded asset is delivered to many viewers. The encoding cost is paid once; the CDN savings scale with delivery. The break-even is roughly the moment your single-asset audience clears 5–10 k viewer-hours.

VOD with high re-watch. Always wins. Encode once at SVT-AV1 preset 4–6, deliver thousands of times, save 30–50 % on CDN.

Live events with large concurrent audiences. Wins above ~50 k peak concurrent viewers per event. Below that, HEVC is cheaper because live encoding cost dominates.

Long-tail VOD libraries with low view counts. Loses. The encoding cost is amortised across too few deliveries. Stay on HEVC or H.264.

Real-time interactive (sub-second). Loses. AV1 encoding latency at preset 10 is still tight; live AV1 in WebRTC works for some flows but adds risk. Use HEVC or VP9 for low-latency real-time tiers.

Why AV1 wins on bandwidth

AV1 (AOMedia Video 1) was finalised in 2018 by the Alliance for Open Media. The codec uses block sizes up to 128×128, motion-compensated prediction with up to 8 reference frames, switchable transforms (DCT/ADST), constrained directional enhancement filter (CDEF) and a film-grain synthesis tool that reduces bandwidth for grainy content without quality loss.

The numbers (Netflix benchmarks). AV1 streaming sessions use roughly one-third less bandwidth than both AVC (H.264) and HEVC at matched VMAF. AV1 reports 45 % fewer buffering interruptions worldwide. Netflix runs ~30 % of all streams on AV1 in 2026.

The trade-off. Encoding is expensive. SVT-AV1 at preset 4 (high-quality VOD) is roughly 5–10× slower than x265 medium. Hardware AV1 encoders close the gap but quality lags software at the same bitrate. The math therefore depends on the ratio of encode cost to delivery savings.

Device support matrix — 2026 reality

Encoder strategy — SVT-AV1, hardware, presets

SVT-AV1 (software, Intel-led). Production-grade software AV1 encoder, Apache 2.0 licensed. Preset 0 is highest quality, slowest; preset 13 is fastest. For VOD: preset 4–6 is the sweet spot. For live: preset 8–10 with constraint-based bitrate caps.

libaom (Google reference). Reference implementation. Higher quality than SVT-AV1 at matched preset, but 3–5× slower. Use only for highest-tier VOD where encoding cost is amortised across millions of views.

NVIDIA NVENC AV1 (hardware). RTX 40-series and newer. Real-time at 4K 60 fps. Quality slightly below SVT-AV1 preset 8 at the same bitrate, but encode cost drops 100×. Best for live encoding clusters.

Intel QuickSync AV1 (hardware). Intel Arc and 12th-gen Xeon. Cheaper per-stream than NVIDIA on commodity hardware. Quality competitive at high bitrates, falls behind below 2 Mbps.

Apple ProRes / Apple AV1 (Silicon). M3 GPUs decode AV1; M4 Pro/Max encode. Promising for studio workflows, less relevant for server-side encoding clusters.

The cost math — when AV1 break-evens

Worked example 1: VOD platform, 10k DAU, 1.5h average daily viewing. Total bandwidth before AV1: roughly 14 TB/day at 5 Mbps H.264. AV1 cuts to ~8 TB/day. CDN cost @ $0.025/GB: $11.5k/month savings. SVT-AV1 encoding for the daily catalogue at preset 4: roughly $1.5k/month additional compute. Net savings: $10k/month.

Worked example 2: live event, 100k peak concurrent, 1.5-hour event. Bandwidth: ~57 TB delivered. AV1 saves ~32 TB. CDN cost saving @ $0.018/GB (volume tier): ~$575. Encoding cost via NVENC: identical to HEVC (real-time hardware). Net: $575 saved per event. With 50 events/year: $28k savings, modest.

Worked example 3: long-tail VOD, 300 hours of content, 200 average views per asset. Encoding cost SVT-AV1 preset 4 @ ~$1.50/h: $450. CDN savings: $300/month. Break-even at month 1.5. Beyond that, pure profit.

The break-even formula. AV1 wins when: (bytes saved per viewer) × (CDN price/byte) × (viewers per asset) > (encoding cost premium). For VOD that ratio is almost always positive. For live with under 50 k concurrent viewers per event it is often marginal.

The multi-codec ladder — the smart approach

Most production stacks do not pick “all AV1” or “all HEVC.” They run a multi-codec ladder: each variant in the manifest tagged with codec + resolution + bitrate, the player picks the best the device supports.

Manifest-level codec selection. HLS supports separate variant playlists per codec; CMAF DASH the same. The player’s codec capability negotiation picks AV1 when supported, falls back to HEVC, then H.264.

The storage trade-off. Triple-codec ladder roughly triples storage cost. For most CDNs and origin storage, this is small compared to delivery savings: storage is $0.02/GB/month, delivery is $0.02–$0.05/GB/delivery. Storage triples, but a popular asset is delivered hundreds of times.

Per-title encoding. Netflix-style adaptive ABR: each title gets its own ladder based on content complexity. Cartoons need fewer rungs; sports need more. AV1 amplifies this benefit because the bitrate savings vary more with content.

Live AV1 challenges

Encoding latency. SVT-AV1 preset 10 hits real-time at 1080p; preset 8 needs hardware acceleration to keep up. Buffer: 2 GOP minimum, ~2 s added to glass-to-glass.

Multi-bitrate live ladder. Encoding 5 rungs of AV1 in parallel needs serious compute. NVIDIA NVENC AV1 at $0.50/hour per stream is the cost-effective answer; software live AV1 across 5 rungs costs $1.50–$3/hour per channel.

WebRTC AV1 support. Chrome, Edge and Firefox decode AV1 in WebRTC. Most browsers do not yet AV1-encode in WebRTC (will come in 2026–2027). For real-time interactive flows, HEVC or VP9 remain the safer pick.

WHIP/WHEP + AV1. Both protocols are codec-agnostic; AV1 in WHIP/WHEP pipelines works in 2026 if both encoder and player support it. Test thoroughly before production.

VOD pipeline — the easy win

A VOD AV1 pipeline is the highest-confidence migration. Workflow:

1. Source ingest → mezzanine. Lossy or visually lossless mezzanine (ProRes, JPEG XS, or master MP4 H.264 high bitrate).

2. Per-title analysis. Optional CRF-tuning step that picks the bitrate ladder per asset complexity.

3. Multi-codec encoding. SVT-AV1 preset 4 for AV1, x265 medium for HEVC, x264 medium for H.264. Five rungs each: 240p, 480p, 720p, 1080p, 4K.

4. Manifest packaging. CMAF with multiple variant playlists per codec. Subtitles, audio tracks fanned out separately.

5. CDN distribution. Origin shield, regional caches, signed URLs for paywalled content.

Vendor matrix — AWS, Bitmovin, Mux, Theta, Dolby

AWS Elemental MediaConvert + AV1. Live AV1 encode, on-demand AV1 with custom presets. AWS pricing per minute encoded; for live, $0.60–$1.20/hour for 1080p output, less for hardware-accelerated.

Bitmovin. Multi-codec encoder service with strong AV1 support. Per-title encoding optimisation, ABR ladder generation. Higher per-minute cost than AWS but better quality control out of the box.

Mux. Developer-friendly video API with auto-generated codec ladders. Premium tier includes AV1. Best when you want the API simple and the underlying complexity hidden.

Theta Labs. Decentralised CDN with AV1 support, pitched at cost-sensitive OTT. Smaller catalogue of integrations than AWS or Mux, but per-GB cheaper.

Dolby OptiView (formerly Hybrik / Dolby.io). Enterprise transcoding with AV1, plus low-latency streaming. Strong on broadcaster-grade SLAs.

Mini case — AV1 saved an OTT $42k/month

An OTT client (NDA, similar profile to Mangomolo) running a VOD library of ~8 000 hours, 45 k DAU, was burning $145k/month on CDN delivery in late 2024. The split: 70 % H.264, 30 % HEVC, no AV1.

The 14-week project. Weeks 1–2: device-share audit (88 % of delivered minutes hit AV1-capable devices). Weeks 3–5: SVT-AV1 preset 5 encoding cluster on Intel Arc QuickSync hardware in their AWS VPC. Weeks 6–9: re-encoded the top 1500 most-watched assets (covered 92 % of delivery), kept H.264 + HEVC for legacy. Weeks 10–12: manifest changes, player codec preference logic. Weeks 13–14: gradual rollout, A/B verification.

Outcome. CDN bill dropped from $145k to $103k/month — $42k/month savings. Encoding cost added $4.5k/month. Net savings: $37.5k/month. Project paid for itself in 4 months. Buffering rate dropped 38 % on AV1-capable devices because of better bitrate efficiency. Want a similar audit? Book a 30-min call.

A decision framework — adopt AV1 in five questions

Q1. What is your CDN bill as a percentage of revenue? >15 %: AV1 is mandatory to chase. 5–15 %: AV1 worth a 14-week project. <5 %: deprioritise.

Q2. Are your viewers on devices that support AV1? Run a 2-week audit on user-agent + reported codec capabilities. Need 60 %+ for the dual-codec path to make sense.

Q3. VOD or live? VOD: high confidence, ship now. Live with >50 k concurrent: yes, with hardware encoder. Live with smaller audiences: defer.

Q4. Encoder strategy? Cloud transcoding service (AWS, Bitmovin, Mux): fastest. Self-hosted Intel Arc / NVIDIA cluster: cheapest at scale. Pure software CPU: only for highest-tier VOD.

Q5. Storage budget? Triple-codec storage requires headroom. If origin storage is cost-constrained, dual-codec (HEVC + AV1) hits 95 % of the savings at lower storage cost.

Pitfalls to avoid

1. Encoding live AV1 on CPU. SVT-AV1 preset 10 needs serious cores. A single 1080p live stream eats 16+ cores; multi-bitrate ladder eats 64+. Use NVENC AV1 or QuickSync hardware.

2. Skipping per-title analysis. Generic ABR ladders waste bandwidth on simple content (cartoons, talking heads) and starve complex content (sports, action). Per-title encoding amplifies AV1’s benefit.

3. Forgetting subtitle and audio tracks. Multi-codec ladder doubles the manifest plumbing. Subtitles (WebVTT) are codec-agnostic but the manifest packaging often missed in initial builds.

4. Trusting browser codec capability strings. Chrome reports AV1 capable but old GPUs do software-only decode at high battery cost. Implement quality-aware fallback.

5. Not testing CTV. Smart TV codec support is uneven; some 2022 TVs claim AV1 but stutter at 4K. Always run a 2-week QA pass on representative panels before flipping the manifest.

KPIs to measure

Quality KPIs. VMAF score on AV1 vs HEVC at matched bitrate (target: matched or +1–2 points). Buffering rate on AV1-capable vs H.264 viewers (target: -20 % on AV1). Decode failure rate (target: < 0.1 %).

Business KPIs. CDN cost per delivered minute (target: -25–40 % post-AV1). Storage cost per asset (target: 2.5–3× pre-AV1). Net cost savings monthly. Time-to-payback on encoding investment.

Reliability KPIs. Encoder pipeline uptime (target: 99.95 %). Manifest generation success rate (target: 100 %). Player fallback rate from AV1 to HEVC (target: < 8 %; high values mean AV1 manifests broken).

When NOT to migrate

Low-traffic VOD. Below 100k delivered minutes/month, encoding cost dominates and AV1 loses. Stay on HEVC or H.264.

Real-time interactive (sub-second). AV1 encoder latency is still tight for sub-1 s glass-to-glass. HEVC or VP9 is safer for live betting, watch parties, conferencing.

Audience dominated by legacy hardware. If 70 %+ of your delivered minutes go to non-AV1 devices, the encoding investment cannot be amortised.

Strict CDN tier with no per-GB savings. Some flat-rate CDN deals do not pass bandwidth savings to you. Check the contract.

FAQ

What to Read Next

Ready to chase 30% off your CDN bill?

AV1 is at the inflection point in 2026 — device support cleared 88 %, hardware encoders are mature, Netflix runs 30 % of streams on it. For VOD with re-watch, the migration pays back in months. For live with >50 k concurrent, hardware AV1 wins. For long-tail VOD or sub-second real-time, hold for now.

The smart pattern is multi-codec: AV1 for capable devices, HEVC for the middle, H.264 fallback. Storage triples but delivery savings dominate any reasonable traffic profile. Run a 2-week device audit, build a per-title encoding pipeline, and ship the manifest changes one cohort at a time.