Bluetooth Channel Sounding vs UWB: Splitting the Warehouse Tag Fleet

What accuracy do Channel Sounding and UWB actually deliver in racking?

Physics sets the floor. UWB under IEEE 802.15.4z uses 499.2 MHz channels, which gives a raw time resolution near 30 cm and lets a receiver separate the direct path from a rack reflection arriving a few nanoseconds later. In live warehouses, two-way ranging against fixed anchors holds 10 to 30 cm, including partial non-line-of-sight behind racking.

Bluetooth Channel Sounding, standardized in Core 6.0 in September 2024, measures phase across roughly 72 channels in the 2.4 GHz band, about 80 MHz of total span. That bandwidth only separates reflections more than about 2 meters apart in path length, so everything closer must be untangled in software. Vendor line-of-sight demos show 10 to 50 cm. In dense racking with metal, conveyor, and forklift traffic, expect:

• UWB: 10 to 30 cm, degrading gracefully under obstruction
• Channel Sounding: 0.5 to 1 m typical, with occasional multi-meter outliers in blocked aisles

Plan on a meter, not thirty centimeters.

What does a tag really cost in 2026?

A Channel Sounding tag is a plain BLE tag with newer silicon: one SoC (Nordic's nRF54 series is the reference example), one 2.4 GHz antenna, one coin cell. As that silicon reaches volume through 2026, finished tags land in the $5 to $12 range at 10,000-unit quantities, close to today's BLE beacon pricing.

A UWB tag carries two radios: the UWB transceiver (Qorvo DW3000 family or NXP Trimension class) plus a BLE SoC for provisioning and wake-up, two antennas, and a more demanding RF layout. Finished tags run $15 to $40 depending on housing and battery.

Power favors Channel Sounding at low duty cycles too. A narrowband ranging exchange draws less peak current than a UWB burst, so a tag reporting every few minutes lasts longer on the same coin cell.

At 20,000 pallets, a $15 per-tag gap is $300,000 before you buy a single anchor.

The tag gap is roughly three to one.

Which assets belong on Channel Sounding?

Count your tags by tolerance. Pallets, totes, returnable containers, rolling cages, and shared tools make up the bulk of most warehouse fleets, often 10,000 to 100,000 tagged items against a few dozen vehicles. The questions those tags answer are aisle-level: which zone holds the pallet, how long has this tote dwelled at pack-out, did the container leave through dock door 4 or dock door 5. A 1 meter answer resolves all of them, and a 2 meter search radius still beats twenty minutes of walking the floor.

These are also exactly the assets where per-tag price decides the business case, because the count multiplies every dollar. They gain phone readability too: Android's Ranging API exposes Channel Sounding alongside UWB, so as supported handhelds spread through 2026 device cycles, a picker can range to a tag without dedicated reader hardware.

High counts, loose tolerance, Channel Sounding wins.

Which assets should stay on UWB?

Keep UWB wherever the error budget is tighter than the spacing between possible answers. Adjacent pallet positions in a rack bay sit roughly a meter apart center to center, so a 1 meter error puts the pallet in the wrong slot and corrupts putaway verification. That workload needs UWB's 10 to 30 cm.

Vehicles are the other hard case. Forklift collision warning and AMR traffic management need position at several updates per second with sub-100 ms latency, under fast motion, near large moving metal. Channel Sounding has not demonstrated that envelope in production. The economics also flip: a site runs dozens of vehicles, not tens of thousands, and the tags draw from vehicle power, so a $40 tag or a $200 vehicle unit is noise against a $40,000 forklift.

• Stay UWB: forklifts, AMRs, pallet jacks, slot-level putaway checks, high-value work in process

If the slot matters, keep UWB.

When do you switch, and in what order?

Treat 2026 as the year to split purchasing, not to rip anything out. Channel Sounding silicon is shipping, but the catalog of tags and gateways is still thin next to a decade of UWB products. Three conditions before moving a tag class:

• Tolerance for that class is 1 meter or looser, verified in your racking, not a demo hall
• Tag count is high enough that the roughly three-to-one tag price gap funds new gateways, because legacy BLE RSSI gateways cannot perform Channel Sounding ranging
• Your positioning platform fuses both radios into one coordinate system, so zone definitions stay consistent during the transition

Then switch at natural replacement points: new tag purchases go to Channel Sounding, existing UWB tags live out their batteries, and vehicle hardware stays untouched. Budget 12 to 24 months of dual-radio operation rather than a cutover weekend.

Switch the fleet, not the forklifts.

Does switching radios reset your patent exposure?

Both radios land on the same patented ground. Two-way ranging protocols, multipath mitigation, anchor calibration, motion-model smoothing, and fusion of RF with inertial or vision data were patented well before Bluetooth 6.0 shipped, and the claims typically read on the method, not the radio. Teams that change radios to cut the BOM sometimes assume the IP question resets with the silicon. It does not.

Position Imaging licenses hundreds of granted US patents in RF ranging, real-time positioning, and sensor fusion, cited by Apple, Bosch, and other major firms, including US 11,774,249, US 12,079,006, US 12,066,561, and US 12,000,947. A license buys two things at once: freedom to operate across both radio choices, and ranging and fusion techniques already proven in production, turning a build measured in years into an integration measured in months.

The radio changes, the ranging math does not.