Apple's UWB APIs vs BLE Channel Sounding: Which Stack for 2026?

What do Apple's UWB APIs give an asset tracker in 2026?

Apple has been widening developer access to its UWB hardware since 2020, and the pieces now add up to a usable asset-tracking stack. Nearby Interaction shipped in iOS 14 for phone-to-phone ranging. iOS 15 opened it to third-party accessories, so a tag built on an interoperable UWB chip (Qorvo DW3xxx or NXP Trimension parts are the common choices) can range against any iPhone 11 or later. iOS 16 added background sessions, meaning a paired accessory keeps ranging while the app is off screen, which is what hands-free workflows like curbside pickup and dock-door handoff actually need. On hardware, Apple's second-generation UWB chip (iPhone 15, 2023) extended Precision Finding range roughly 3x over the original U1. What you get as a builder: distance to about plus or minus 10 cm, direction angles when the accessory sits in the camera's field of view, all over standard 802.15.4z HRP UWB on channels 5 and 9. The session bootstraps over BLE, then ranging runs on UWB. Every iPhone since 2019 is free infrastructure.

Where does Bluetooth 6.0 channel sounding actually stand?

Channel sounding is real now. The Bluetooth SIG published it in Core 6.0 in September 2024, combining phase-based ranging hopped across up to 72 channels in the 2.4 GHz band with a round-trip-time measurement to defeat distance spoofing. Silicon is shipping in volume: Nordic's nRF54L15 and NXP's KW47 both support it, and Android 16 exposes it alongside UWB through a single ranging API, so the Android side of the ecosystem is genuinely in place. What is not in place is accuracy parity or iPhone support. Vendor demos in clean line of sight show 10 to 50 cm. In a racked warehouse or a hospital corridor, with metal, bodies, and carts generating multipath, reported results cluster at 1 to 3 m. And no iPhone exposes channel sounding ranging to developers; Apple's precision-location features, AirTag and Precision Finding included, are built on UWB, and Apple has published no channel sounding API so far. Great silicon, strong Android story, meter-class reality.

Why does 500 MHz beat 80 MHz indoors?

Indoor ranging accuracy is mostly a multipath problem, and multipath resolution is set by bandwidth. The radar math: range resolution is c divided by twice the bandwidth. A 500 MHz UWB channel separates signal paths about 30 cm apart, so the receiver can pick the direct path's leading edge out of the reflections and deliver 10 to 30 cm fixes. Channel sounding synthesizes roughly 80 MHz by hopping 1 MHz channels, which separates paths only about 1.9 m apart. Indoors, the direct path and the first strong reflection usually arrive closer together than that, so they merge and the phase-based estimate lands somewhere between them. That is the bias behind the 1 to 3 m field numbers, and no firmware update removes it, because the limit is the channel, not the radio. Clever estimators, super-resolution and ML-based first-path pickers, claw back margin in benign rooms and give it back in dense ones. Bandwidth is the accuracy budget. Physics does not negotiate.

Which stack should you pick? Start with the phone

• Someone ranges to your tag with their own phone: build UWB. Nearby Interaction is the only sub-meter ranging path on iOS, and it covers every iPhone back to 2019. A picker finding one pallet among 40 with a handheld, a shopper locating a curbside order, a nurse finding an IV pump: all of these fail at 2 m accuracy and work at 20 cm.
• Closed facility, dedicated anchors, zone-level answers: channel sounding earns a look. If 'which bay, which room' is the real question, 1 m is enough, and a single-radio 2.4 GHz tag cuts BOM while one radio handles ranging, telemetry, and firmware updates.
• Both tiers in one building: run them together. Channel sounding covers the long tail of low-value assets, UWB covers the spots where sub-meter pays, and one location engine fuses both.

The failure mode to avoid is picking the cheaper radio, then discovering eighteen months later that the use case was a phone use case all along. If an iPhone is in the loop, build UWB.

What does it take to ship either stack well?

The radio is the easy part. The hard part is identical for both stacks: first-path detection under non-line-of-sight, anchor time synchronization, motion models that fuse ranging with inertial data, and tag power budgets that survive a 3-year battery claim. Teams that build this from scratch typically spend 12 to 24 months before accuracy holds up outside the lab, and they do it inside a patent field that has been staked for over a decade. That is the case for licensing instead. Position Imaging licenses hundreds of granted patents across RF ranging, real-time positioning, and the computer vision and machine learning methods that clean up ranging data, including US 11,774,249, US 12,079,006, US 12,066,561, and US 12,000,947. The portfolio is cited by Apple and Bosch, which tells you where the prior art in this space actually sits. Licensing puts your engineers on the product layer instead of re-deriving first-path estimation, and it gives your counsel a known cost instead of a freedom-to-operate guess. License the foundations. Ship the product.