In late January, Frankfurt Airport started piloting Auracast, pushing every gate announcement at two of its gates straight into passengers’ hearing aids, earbuds, and phones, and GN, the Danish hearing group behind the pilot, called it the first airport in the world to do so. Eight months later, on September 9, Apple launched the iPhone 18 family, the foldable iPhone Duo and AirPods 5. Not one of them can receive that broadcast on its own.
That gap is the Auracast story in 2026. The radio design was finished years ago; the LE Audio silicon underneath it has shipped in hundreds of millions of devices; every major hearing aid maker was on board by August; and TVs from Samsung, LG, Panasonic and Philips now transmit it. The phone that decides what the US market considers normal still acts like the feature doesn’t exist. I’ve been close to mobile Bluetooth silicon since I started working with Qualcomm Snapdragon in early 2006, and I struggle to name another Bluetooth feature this useful that took this long to reach ordinary people. I’ve only used Auracast once myself, at a tech demo, which is the friendliest room it will ever play in; a crowded gate at Frankfurt is the version I’d like to stand in.
Classic Bluetooth was built for one listener
A2DP, the profile that has carried your music since the mid-2000s, is a point-to-point link between one source and one sink, with SBC doing the compression and the sink acknowledging packets as they arrive. Everything the industry built on top of it has been a workaround. True wireless earbuds needed either a relay, where the primary bud receives the stream and forwards it to the other one, or a sniffing trick where the second bud eavesdrops on the link, which is roughly what Qualcomm’s TrueWireless Mirroring does. Sharing with a friend meant a second full link: Samsung’s Dual Audio on the Galaxy S8 in 2017 simply ran two A2DP connections side by side, and Apple’s Audio Sharing only works between AirPods and Beats.
Every extra listener in that model costs another connection, another slice of airtime, and more drain on the phone’s radio, which is why nobody ever tried to scale it to a departure gate. Broadcast flips the relationship. The transmitter sends one stream into the air and neither knows nor cares whether two people are listening or two thousand.
What actually happens on the air
Auracast is the consumer brand for the broadcast half of LE Audio, and LE Audio rests on one feature added to the Bluetooth Core Specification 5.2, adopted at the end of 2019 and shown off at CES in January 2020: isochronous channels. Isochronous data has an expiry date. Each audio frame has a window in which it’s useful, and once that window closes, the radio stops trying to deliver it and moves on. That sounds like a downgrade until you realize it’s what makes synchronization possible. Every receiver renders a given frame at the same agreed moment, set by a presentation delay the source announces, so your left and right buds stay locked together, and a whole room hears the same syllable at the same instant.
The spec splits isochronous traffic into two shapes. Connected Isochronous Streams (CIS) carry unicast audio to a specific device with acknowledgments, and they replace A2DP inside LE Audio earbuds. Broadcast Isochronous Streams (BIS) are grouped into a Broadcast Isochronous Group (BIG) and sent to nobody in particular. A broadcast has no acknowledgments at all, because a transmitter serving an unknown number of receivers can’t wait for replies. Reliability comes from repetition instead. The BIG parameters set how many times each packet goes out and when: a burst number, a number of subevents per interval, an immediate repetition count, and a pre-transmission offset that lets the source send some packets early so a receiver that misses one copy gets another shot. It’s a blunt trade-off of airtime for robustness, and when every copy of a frame is lost anyway, LC3’s packet loss concealment papers over the hole.
Finding a broadcast is where most of the cleverness lives, and it took me a while to appreciate how layered it is. A source first announces itself on extended advertising, carrying a 24-bit Broadcast ID and a human-readable Broadcast Name like “Gate B42” or “Screen 3”. That advertisement points to a periodic advertising train, a steady beacon on a fixed schedule, and the train carries two things. One is the BASE, the Broadcast Audio Source Endpoint structure, which describes the subgroups, the codec settings, which BIS carries which audio channel, and metadata such as language. The other is the BIGInfo, which gives a receiver the exact timing it needs to lock onto the BIG and says whether the streams are encrypted.
The BASE is why one transmitter can carry more than one thing. A conference hall can put English and French in separate subgroups, and a TV can send left and right as two BISes so each earbud syncs only to the channel it needs. The receiver reads the BASE, picks the stream it wants, and ignores the rest.
Privacy is a single switch with a long key. A broadcast is either public or protected by a 16-octet Broadcast Code, and GN’s Auracast+ app for its TV-Streamer+ lets you flip between an open broadcast and a password-protected one, which is exactly the knob you want at home so your neighbor’s earbuds don’t pick up your late-night movie.
What makes Auracast usable rather than theoretical is the Broadcast Assistant, and it’s the piece I’d explain first to anyone building a product. Earbuds and hearing aids are terrible scanners. They have tiny batteries, no screen, and no way to show you a list of eleven nearby streams, so the spec lets another device do the looking. A phone, a watch, or a hearing aid’s companion app acts as the Assistant: it scans, shows you the list, and when you tap one it connects to your buds over GATT, writes the chosen source into the Broadcast Audio Scan Service running on them, and hands over the periodic advertising sync using PAST, a Bluetooth 5.1 feature that lets one device pass its timing lock to another so the buds don’t have to hunt for it. If the stream is encrypted, the Assistant can deliver the Broadcast Code in the same exchange.
This is also why iPhone owners with the right hearing aids aren’t locked out. The hearing aid does the receiving itself, and the phone only runs the manufacturer’s app to pick the stream; ReSound’s app has worked that way on iPhones for a while. It’s a clever workaround, and a slightly humiliating one for a company that sells itself on accessibility.
Venues got one more shortcut more recently with the Broadcast Audio URI spec, which packs a broadcast’s name, ID, and code into a QR code you scan with your phone. The Public Broadcast Profile adds a small flag to the advertisement saying whether a stream is standard quality (16 or 24 kHz sampling) or high quality (48 kHz), so a receiver knows what it’s getting before it joins.
The audio itself is LC3, the Low Complexity Communication Codec that Fraunhofer IIS and Ericsson developed for LE Audio. It works on 10 ms frames, with a 7.5 ms option, and the Bluetooth SIG’s listening tests put it ahead of SBC at roughly half the bitrate. The broadcast presets tell you what Auracast is designed for. A speech stream at 16 kHz and 32 kbps is plenty for a gate announcement or a sermon, and the common 48 kHz music setting runs at 80 kbps per channel. That’s clean, perfectly good audio, but it isn’t a hi-res pipe, and anyone selling Auracast as an audiophile feature is selling the wrong thing.
The hardware, and why the version number lies
On paper, the requirements are short. A source needs a Bluetooth 5.2 or later controller with isochronous support, a host stack that implements the Broadcast Source role from the Basic Audio Profile, and an LC3 encoder. A receiver needs the same controller generation plus the ability to sync to a BIG and decode LC3, and ideally the Scan Delegator role so an Assistant can steer it.
In practice, the number printed on the box tells you almost nothing. I made the same complaint in my Nothing Ear (3a) review two weeks ago: a version number doesn’t bring LE Audio or Auracast with it, and I’m still surprised how many spec sheets act as if it does. Bluetooth 5.3 or 6.0 on a spec sheet doesn’t mean LE Audio, and LE Audio doesn’t mean Auracast is switched on; these features are optional, and the firmware and the operating system both have to expose them. Google’s own lineup shows how uneven this gets: Android 16 brought Auracast to the Pixel 8 and later, but left out the cheaper Pixel 8a and 9a, even though both support LE Audio and list Bluetooth 5.3. Nothing’s Ear (3a), which I reviewed in September, ships with the silicon but didn’t expose LE Audio or Auracast at launch.
Frustrating, and completely normal.
On the chip side, the earbud and headphone market is well served. Qualcomm’s Snapdragon Sound platforms have supported LE Audio and Auracast for several generations, and the Sound Elite Gen 2 it launched at Snapdragon Summit in September integrates Bluetooth and micro-power Wi-Fi 6E on the same audio platform for the first time, which I’ll come back to. Airoha, Bestechnic and Realtek supply LE Audio parts to much of the rest of the TWS market, while Nordic’s nRF5340 has been the default board for engineers prototyping LE Audio since the early days. You can see the result on store shelves: my sub-$100 earbud comparison had a budget pair listing Auracast next to LDAC, and Nothing’s new Headphone (1) Pro can act as a source itself, broadcasting whatever you’re hearing to other people’s headphones.
Venue hardware is its own category. Ampetronic and Listen Technologies, now operating as one company, sell the Auri system: TX2N wall or ceiling transmitters that take analog or Dante audio, run on PoE and are managed over the network, plus RX1 receivers and neck loops for people whose hearing aids only have a telecoil. At the cheap end, a USB dongle like the FlooGoo FMA120 turns a laptop into a transmitter. A growing share of transmitters in people’s homes are televisions, with Samsung, LG and Panasonic already shipping Auracast TVs and Philips adding it to much of its 2026 range from May.
Hearing loss is the killer app, whatever the marketing says
The Bluetooth SIG markets Auracast with friends sharing a playlist on a bus, and that use case is real but small. The market actually moving money is assistive listening. For decades, the answer for hearing aid wearers in a theater or a ticket hall was the induction loop: copper cable run around the room, a telecoil in the hearing aid, and an installation bill that pushed plenty of venues to settle for a sticker suggesting a loop might be working. Auracast lets one transmitter serve hearing aids, cochlear implant processors, and everyone else’s earbuds at the same time, in stereo and on multiple channels, and the listener brings the receiver.
The hearing aid industry took its time. Phonak held out longest, and its critics accused it of protecting its own proprietary streaming accessories. By August 2026, that argument was over: ReSound launched the Sensia, Oticon the Reveal, and with Phonak and Signia shipping too, every major brand now supports Auracast. That’s the moment I’d mark as the tipping point, more than any phone launch, because a hearing aid wearer who suddenly hears a gate announcement clearly becomes the most persuasive salesperson this standard will ever have.
Deployments are spreading unevenly, and mostly in Europe. Frankfurt is testing it as part of a “silent airport” approach, where constant loudspeaker calls give way to direct personal updates. In the UK, the charity RNID used its Clear Signals research in September to call for Auracast across public transport, after finding that inaccessible announcements leave millions of travelers anxious. The Basilique Notre-Dame-de-la-Victoire in Saint-Raphaël became the first basilica in France to install it, which made the Frenchman in me smile, and rail suppliers were demonstrating it at InnoTrans in Berlin in September. Classrooms are next, with SMART Technologies building Auracast into its SoundFlow classroom audio system.
The consumer side leans on TVs and speakers. A gym or sports bar can unmute every silent screen by giving each TV its own broadcast name. Party speakers are an easy win: Sony just announced three new ULT POWER SOUND models with Auracast, and JBL has used it to link its portable speakers since 2024. Museums and tour operators get to stop renting and disinfecting audio guides, a line item they’ll notice faster than any accessibility argument.
The venue business case is less clean than the brochures suggest, and I’d push back on anyone promising a zero-hardware accessibility budget. In the US, as I understand the ADA rules, assembly venues still have to provide assistive listening receivers, including hearing-aid-compatible ones, for a share of seats, so “bring your own earbuds” supplements that obligation rather than replacing it, and hearing-industry groups are still working out standards for where Auracast fits. The savings show up in running the thing: a PoE transmitter on the network can be reconfigured from a laptop, and most receivers are phones and earbuds people already own and keep charged.
For chipmakers, the prize is volume. ABI Research’s Auracast retrofit white paper puts annual shipments of LE Audio devices at about 3 billion by 2028, up from 391 million in 2023. Once LE Audio is standard in every mid-range earbud, Auracast stops being a feature anyone can charge for and becomes a checkbox, the same fate multipoint met.
After 2 Mbps
The history so far is short and lumpy. LE Audio was announced at CES 2020; the full specification set and the Auracast name only arrived in the summer of 2022, and the first wave of products spent the next two years with the hardware in place and the software missing. Samsung moved first, pairing Galaxy phones and TVs with its own Buds, and Google made Auracast a system feature in Android 16 last year. Consumers barely noticed until hearing aids and TVs dragged it into public spaces over 2025 and 2026.
The next jump is in the radio itself. The Bluetooth SIG expects to adopt High Data Throughput in a Core Specification release late this year, lifting Bluetooth LE’s top rate from 2 Mbps to as much as 7.5 Mbps with higher-order modulation and forward error correction at the physical layer. It’s a hardware change, so nothing you own today gets faster. Work on moving Bluetooth LE into the 5 GHz and 6 GHz bands is running alongside it, as is work on lossless and high-resolution audio.
I’m less excited about what HDT does for Auracast than the press releases imply. Broadcast has to serve the weakest receiver in the room, usually a hearing aid with a battery the size of a lentil, so the streams that matter most will stay modest LC3 for years. HDT matters far more for unicast earbuds, and there Qualcomm isn’t waiting: Sound Elite Gen 2 already puts micro-power Wi-Fi next to the Bluetooth radio so earbuds can reach cloud services directly instead of relaying every request through the phone, which takes the heaviest traffic off Bluetooth altogether. A proprietary lane running next to an open one is the tension I’ll be watching once the SIG’s spec lands.
Which brings me back to Cupertino. The iPhone 18 Pro lists Bluetooth 6, which proves nothing on its own for all the reasons above; what’s holding Apple back is its own software stack and product choices. That stings, because Apple’s accessibility work, from turning AirPods Pro into hearing aids to putting Live Translation across the AirPods lineup, targets exactly the people Auracast was built for. Hearing-aid forums have predicted Apple support every year since iOS 17, and every September has ended the same way, so I’ve stopped predicting. I suspect US venues are looking at the iPhone’s share of their audience and holding off while Europe installs transmitters in airports and basilicas. Auracast wins either way, because hearing aids and TVs have already made it ordinary. Apple only gets to decide whether American venues catch up in 2027 or somewhere around 2030, and right now it seems perfectly happy to let that clock run.