IN Brief:
- Nordic has demonstrated Bluetooth High Data Throughput on a prototype system-on-chip.
- Proposed HDT physical layers extend the nominal radio rate from 2 Mbps to as much as 7.5 Mbps.
- Bluetooth SIG release is expected later in 2026; existing Bluetooth LE chips cannot gain HDT through software alone.
Nordic Semiconductor has demonstrated Bluetooth High Data Throughput (HDT) operating on prototype silicon, showing hardware support for a proposed enhancement to Bluetooth Low Energy communications. The technology is intended to raise the maximum physical layer data rate from the current 2 Mbps to as much as 7.5 Mbps while retaining Bluetooth LE as the underlying wireless ecosystem.
Nordic’s prototype operates ahead of the Bluetooth Special Interest Group’s planned completion of the HDT specification later in 2026. Future devices will need compatible new radio silicon, but Nordic intends developers to keep using much of its existing Bluetooth LE software and tooling. The demonstration therefore establishes an implementation path while the formal standard and commercial product specifications are still being completed.
To increase the amount of data conveyed over the radio channel, HDT adds modulation options beyond the Gaussian frequency shift keying used by Bluetooth LE 1M and 2M. Higher order schemes including π/4-QPSK, 8-PSK and 16-QAM carry more information per symbol, alongside forward error correction. The added capacity also makes receiver sensitivity, interference and signal quality more important to sustaining the fastest modes.
Five proposed physical layers — HDT2, HDT3, HDT4, HDT6 and HDT7.5 — offer different rate and link margin combinations. A device operating under stronger signal conditions may use a faster mode, whereas a weaker or noisier link may need a lower rate to maintain reliability. Link quality feedback is intended to guide those changes rather than require the highest rate throughout a connection.
Application throughput depends on packet airtime and protocol overhead as well as the nominal radio rate. Each packet requires airtime for a preamble and header, while acknowledgements and spacing between transactions add further overhead. HDT supports longer packets that spread these costs across more payload data, increasing bulk transfer efficiency under suitable link conditions. Nordic’s file transfer measurements reflect those overheads alongside the available physical layer rate.
The performance difference depends heavily on payload size. In Nordic’s preliminary desk demonstration, two prototype SoCs transferred a 2 MB file with 508-byte payloads. The fastest HDT mode averaged approximately 3.8 Mbps and completed the transfer in about four seconds, against 1.4 Mbps and 12 seconds on LE 2M. The 7.5 Mbps designation refers to a physical layer rate; actual file transfer throughput is lower after protocol overhead. Very short messages can perform better with the established LE modes.
Completing a data transfer more quickly can shorten the period for which a radio must remain actively transmitting or receiving, potentially leaving more time for other traffic in the shared 2.4 GHz band. Whether a battery powered product actually consumes less energy for a given task also depends on the modulation selected, signal strength, retransmissions, processing overhead and the sleep behaviour of the rest of the device.
Those faster modes cannot be added to existing Bluetooth LE transceivers through a firmware update, because HDT needs different radio hardware. Nordic nevertheless intends the surrounding software development approach and tools to remain familiar on future compatible SoCs. A product moving to HDT would therefore require new silicon and radio validation, even where its application software can build on an established Bluetooth LE design.
Industrial sensors that move sizeable records, medical or hearing devices carrying richer data, and embedded equipment receiving software updates are potential uses for the higher transfer capacity. Short control messages have different packet efficiency characteristics and may continue to suit established Bluetooth LE modes. Application benefits will depend on payload length, link conditions and how frequently each device needs to exchange data.
Nordic is due to show the prototype at electronica 2026 while developing a commercial HDT-capable SoC. Its measured 2 MB transfer demonstrates how packet length and modulation combine to raise useful throughput, with the 7.5 Mbps figure remaining a nominal physical layer rate. Subsequent product specifications will depend on completion of the HDT standard and qualification of the new radio modes.
