The Future of AV1 and Beyond: Why the Codec Wars Are Finally Ending

The Long Road to Royal-Free Video

For the longest time, distributing high-quality video or highly compressed images on the web was a legal minefield. If you used H.264, you owed licensing fees to MPEG LA. If you used H.265 (HEVC), you owed fees to three different patent pools with contradictory overlapping royalties. In an ecosystem built on open standards, the foundational layer of visual media was heavily gated.

Then came the Alliance for Open Media (AOMedia). Backed by everyone from Google and Netflix to Microsoft and Apple, the mandate was simple: construct a world-class, ultra-efficient codec that is entirely open and royalty-free. The result was AV1, and its derivative image format, AVIF.

Now, in 2026, the war is effectively over. AV1 has won. But understanding why it won, and what technologies are poised to follow it, provides a fascinating look into internet infrastructure.

Why AV1 Surmounted the HEVC Monopoly

It is crucial to understand that AV1 didn't win solely because it was "free." It won because it was structurally better suited for the internet delivery model.

1. Superior Compression Efficiency at Low Bitrates

AV1 was designed with streaming logic first. While HEVC performs admirably at 4K Blu-Ray bitrates (20-40 Mbps), AV1's intra-frame prediction logic excels at "starvation bitrates." When Netflix has to push a 720p stream to a rural mobile connection at 800 kbps, AV1 maintains temporal stability (less blockiness) far better than HEVC.

2. The Great Hardware Unification

The major critique of AV1 in its early years (2018-2022) was its computational complexity. Decoding 4K AV1 software effectively melted laptop batteries.

However, the hardware industry synchronized remarkably well. By 2024, AMD's RDNA architectures, Nvidia's Ada Lovelace, Intel's Arc/Meteor Lake, and critically, Apple's M3 and A17 Pro chips all included dedicated hardware AV1 decoders. In 2026, trying to find a newly manufactured smartphone without AV1 decode capabilities is virtually impossible.

3. Apple's Tipping Point

The format wars truly ended the day Apple capitulated. Historically heavily invested in HEVC, Apple's inclusion of AV1 hardware decoding in their silicon and full support in macOS and iOS signaled to the entire web development community that AVIF and AV1 were safe to deploy universally without fear of rendering blank squares on iPhones.

The Image Evolution: AVIF Replacing WebP

AVIF is simply an AV1-encoded intra-frame wrapped in an ISOBMFF container. It is the photographic byproduct of the video wars, and its dominance in 2026 over WebP is profound.

WebP (based on the VP8 video codec) was a massive leap forward from JPEG, but it suffered from limited color depth and poor high-frequency detail retention. AVIF natively supports:

• 10-bit and 12-bit color depth (crucial for HDR displays)
• Lossless alpha channels that rival PNG sizes
• Sub-sampling capabilities that prevent color bleeding in text overlays over photographs.

Our telemetry at Optimage consistently shows AVIF files averaging 30% smaller than equivalent WebP files, a staggering reduction when extrapolated across an entire CDN.

But What Lies Beyond AV1? Enter AV2 and VVC

Technology despises a vacuum. Even as AV1 completes its deployment phase, the successors are already heavily in development.

VVC (Versatile Video Coding / H.266)

Developed by the Joint Video Experts Team (JVET), VVC is the formal successor to HEVC. It promises another 30-50% bitrate reduction over HEVC. It is incredibly efficient, particularly for 8K and 360-degree VR content.

However, VVC suffers from the exact same genetic defect as its predecessor: suffocating patent licensing structures. Despite its technical brilliance, web browsers (notably Chrome and Firefox) remain staunchly opposed to subsidizing patent pools for standard web infrastructure. VVC will likely dominate the broadcast television and physical media realms, but struggle on the open web.

AV2: The AOMedia Sequel

AOMedia isn't resting. AV2 is actively in the research phase. The goal is straightforward: achieve a 30% compression efficiency gain over AV1, while ensuring hardware implementers can reuse significant portions of the AV1 silicon architecture.

Early proposals for AV2 include:

• Enhanced machine learning-based loop filters.
• Better non-local spatial prediction methods.
• More granular block partitioning.

Expect to see finalized bitstreams for AV2 around 2028.

The Role of Neural and AI Codecs

Here is where standard block-based codecs meet their existential threat. Organizations like Meta AI and Google Research are increasingly pivoting away from traditional mathematics-based video encoding.

Instead of defining specific mathematical transforms (like DCT), Neural Image Compression uses autoencoders. An AI model analyzes an image, compresses the latent space features into a tiny payload, and a corresponding AI model on the user's device "hallucinates" the details back upon decoding.

While currently too computationally intensive for widespread real-time decoding on mobile battery, Neural Codecs routinely achieve PSNR scores that outperform AV1 by 40%. As Neural Processing Units (NPUs) become standard hardware in every device by 2027, the line between "video compression" and "AI generative upscaling" will blur completely.

Practical Architecture for Today

While neural codecs and AV2 dominate engineering papers, your architecture today needs stability.

1. Serve AVIF as your primary image source. Use the <picture> tag gracefully falling back to WebP.
2. Utilize hardware encoding. If your backend relies on FFmpeg, ensure you are utilizing av1_nvenc or QuickSync rather than raw CPU encoding with libaom-av1 unless quality per byte is absolutely paramount over speed.
3. Optimize your delivery. Ensure your CDN strips unnecessary metadata natively to squeeze the last 5% out of an AVIF payload.

The codec war is over, and the Open Web finally won. Now, it is time to optimize the peace.