Short answer: three things are changing in IPTV delivery. Low-latency streaming is becoming routine rather than a project, landing most live streams in the 2–6 second range instead of 15–30. AV1 is becoming the default top rung of the encoding ladder as hardware decode reaches near-universal coverage. And over-the-air broadcast is converging on IP through ATSC 3.0, with work underway to bridge it to mobile networks. What isn't changing: licensing, not technology, decides what any service can legally carry.
Everything below about 2027 is a projection from where standards and deployments stood through 2026. The dates are direction, not schedule.
How much latency can you actually expect?
| Protocol | Typical glass-to-glass | Best fit | Main limitation |
|---|---|---|---|
| Standard HLS / DASH | 15–30 sec | Broad-reach VOD and linear | Too slow for live sports alongside broadcast |
| LL-HLS | 2–6 sec (≈1.2–1.8 tuned) | Apple devices, hls.js players | Requires CMAF fMP4 and careful player tuning |
| LL-CMAF over DASH | 3–6 sec | Android, smart TVs, mixed pipelines | Chunked transfer carries more overhead |
| WebRTC | Sub-second | Interactive, two-way, live commerce | Not viable as the primary path on connected TVs |
LL-HLS reaches those numbers through three mechanisms: partial segments, blocking playlist reload, and preload hints, all built on CMAF fragmented MP4 containers. Getting to the low end of the range means reducing chunk size to around 200 ms, tuning the player's target buffer to roughly 0.9–1.2 seconds, and running the packager at the CDN edge.
The practical 2027 architecture for a broad device matrix: LL-HLS as the primary low-latency path for Apple devices and anything using hls.js, LL-CMAF with DASH as the secondary path for Android and smart TVs, and WebRTC reserved for products where sub-second delay is the point. On Roku, Samsung, LG and Google TV, WebRTC is effectively off the table for primary playback — you get the media players those platforms provide.
Worth stating plainly: the last two seconds cost the most. Each increment buys edge infrastructure and player fragility, and a fast stream that stutters is worse than a slow one that doesn't. Measure at the 95th percentile, not the median.
Is AV1 worth switching to yet?
For on-demand, increasingly yes. Netflix measured roughly 48% lower bitrate than H.264 on real content in 2025, and hardware decode is now widespread across phones, TVs and GPUs — which was the objection that kept AV1 sidelined for years.
For real-time WebRTC, H.264 remains the safe default because SFU and browser support is universal; HEVC is the step up from there.
The constraint has inverted. Decoding is solved; encoding is now the expense. AV1 is costly to encode well, which is why it appears first on catalogue content encoded once and watched millions of times, and last on live streams encoded once and watched once. Expect AV1 as the top rung most viewers actually receive by 2027, with H.264 kept alive underneath for a long tail of old devices that never fully disappears.
Why broadcast matters to an IP story
ATSC 3.0 — marketed as NextGen TV in the US — is the first IP-based digital television standard, which puts over-the-air broadcast and internet delivery on the same transport. It also underpins DTV+ in Brazil and UHD Broadcast in South Korea. A memorandum of understanding signed in 2026 between ATSC, Brazil's SBTVD Forum and Korea's TTA aims at interoperability across those implementations.
The development to watch is Broadcast-to-Everything (B2X), an ATSC initiative integrating broadcast and multicast delivery with 3GPP mobile networks. Related work covers addressable multicast — IP multicast over ATSC 3.0 with addressable data flows, so datacasting reaches only intended recipients — with proof-of-concept deployments expected during 2026.
The payoff, if it lands: hybrid delivery, where a device pulls a programme over broadcast spectrum when many people want it simultaneously and over unicast when few do, invisibly. Broadcast is drastically more efficient for mass simultaneous viewing, which is exactly where unicast CDN economics are worst. Live sports is the obvious case.
The brake is regulatory. US broadcasters are waiting on an FCC rulemaking (MB Docket No. 16-142) covering the ATSC 1.0 transition, and that timeline governs how fast capacity can be reallocated.
What's absorbing engineering effort
Multi-CDN orchestration. Single-CDN delivery is now the unusual choice at scale; routing on live quality telemetry is standard.
ML in the pipeline. The real applications are per-title and per-scene bitrate optimisation and anomaly detection in quality monitoring — catching regional degradation before the support tickets arrive.
Timing discipline and skills. IP-native workflows demand rigorous clocking and jitter management, and there's a genuine industry skills gap around PTP, multicast and the associated security controls. This is the quiet constraint on how fast anything above actually ships.
What won't change
Standard-latency HTTP streaming stays the default for most content, because most content doesn't need low latency and cached segments are cheap and reliable. H.264 stays in every ladder. Smart TV fragmentation stays the hardest part of shipping a streaming product.
And the constraint that no protocol addresses: rights determine what a service can carry. Everything above makes delivery cheaper, faster and more efficient. None of it changes what a provider is licensed to distribute. A service advertising tens of thousands of premium channels for a few dollars a month hasn't solved a technical problem — it has skipped a legal one, and the technical improvements described here are irrelevant to that.
Quick answers
Does low latency require WebRTC? No. LL-HLS reaches 2–6 seconds over ordinary CDN infrastructure and works on connected TVs, where WebRTC generally doesn't.
Will AV1 break older devices? Only if you drop the H.264 rung. Keep it in the ladder as a fallback.
Is ATSC 3.0 the same as IPTV? Not the same, but converging. It's IP-based broadcast rather than IP-based unicast delivery, and B2X is the work aimed at joining them.
What's the single biggest limiter in 2027? Device fragmentation and operational discipline, not protocol design.
If you're troubleshooting a stream right now rather than planning for 2027, see Why Your Stream Buffers: A Diagnostic Guide.
Sources: ATSC announcements and SET Expo 2026 sessions on ATSC 3.0 and B2X; TV Tech reporting on FCC MB Docket No. 16-142; Netflix AV1 bitrate measurements (2025); published LL-HLS, CMAF and WebRTC deployment benchmarks from 2025–2026.