Apple's Second-Gen UWB Chip: What It Means for Positioning Startups

What actually changed with Apple's second generation UWB chip?

Apple shipped its first UWB chip, the U1, in the iPhone 11 in September 2019. It implements HRP ultra-wideband per IEEE 802.15.4z on channels 5 and 9, near 6.5 GHz and 8 GHz. The second generation chip arrived four years later in the iPhone 15 lineup and Apple Watch Series 9, and Apple's headline claim was Precision Finding connections at up to three times the range of the U1. For a builder, the range number matters less than the distribution. Every flagship iPhone since 2019 carries UWB, and by 2026 the second generation part has been standard across three iPhone generations. That is centimeter-class ranging hardware your customers already own, at zero bill-of-materials cost to you. Under clean line of sight, UWB time-of-flight ranging lands near 10 cm of distance error, and the iPhone's antenna array adds a direction vector when the peer sits in its field of view. The install base is the real story.

What does the UWB ecosystem look like in 2026?

UWB stopped being an Apple-only bet years ago. The FiRa Consortium runs interoperability certification, and modules built on Qorvo's DW3000 family, NXP's Trimension line, and Murata's integrated parts interoperate with Apple's chips through the Nearby Interaction accessory protocol. On Android, Google exposes ranging through the androidx.core.uwb Jetpack library, but hardware coverage stays thin: Pixel Pro models since the Pixel 6 Pro and Samsung's Plus and Ultra tiers since the Galaxy Note20 Ultra. Automotive pulled the standard forward too. The Car Connectivity Consortium's Digital Key 3.0 specification uses UWB for passive entry, and BMW shipped it as Digital Key Plus back in 2021. Meanwhile IEEE has been finishing the 802.15.4ab amendment, which adds narrowband-assisted modes aimed at lower power and longer range. The radio layer is settled, certified, and multi-vendor. Standards converged; handsets did not.

Is there an AirTag Precision Finding API for third parties?

No. Apple exposes no API that lets a third-party app range against an AirTag. Precision Finding for AirTags lives inside the Find My app, full stop. The Find My network accessory program will put your own tracker into the Find My app, but that solves finding, not positioning, and your app still receives no distance or bearing data from it. The real developer surface is the Nearby Interaction framework. It has supported phone-to-phone ranging since iOS 14, U1-compatible accessories since iOS 15, and background sessions with Bluetooth-paired accessories since iOS 16. The working pattern: embed a certified UWB module in your product, implement Apple's published accessory protocol, and your app receives live distance and direction updates from your own hardware. Anyone searching for an AirTag Precision Finding API is really looking for this path. Build on Nearby Interaction, not AirTags.

Where consumer UWB helps, and where it stops

Consumer UWB is excellent at session-based, user-present tasks:

• Guiding a person to a specific item, dock, or parked vehicle
• Wayfinding to a fixed anchor in a lobby, garage, or store
• Hands-in-pocket access control and proximity-triggered actions
• Pairing and handoff driven by physical approach

It is the wrong foundation for continuous tracking. iOS gives you a ranging session, not a positioning infrastructure. You get a filtered distance and direction vector, not raw channel impulse response or antenna-level angle data, so you cannot run your own multipath rejection. Sessions are foreground-biased, the OS controls duty cycle, and nothing runs while the phone sits in a pocket with your app closed. If the product must track thousands of assets across a warehouse or hospital around the clock, you still need fixed anchors and dedicated tags, and the handset becomes one convenient client among many. The phone is a client, not infrastructure.

How do you build on hardware you do not control?

Treat Apple's UWB stack as a dependency with its own roadmap, because it is one. Apple has already changed the rules once: background ranging arrived in iOS 16 but only for accessories paired over Bluetooth, which forced design revisions on early accessory makers. Four rules keep you safe:

• Make your accuracy spec independent of the handset. Your anchors and tags deliver the guaranteed number; the phone adds convenience on top.
• Fuse UWB with inertial or camera data so a throttled session degrades to coarse guidance instead of failing outright.
• Abstract the ranging layer. The same solver should consume Apple Nearby Interaction, Android androidx.core.uwb, and your own FiRa-certified modules.
• Retest against each iOS beta every summer, since session behavior shifts between releases.

Startups that follow these rules ship features when Apple expands the chip and survive when Apple tightens the API. Own the solver, rent the radio.

Does the patent picture change when Apple owns the radio?

No, and this surprises founders. Apple's patents cover Apple's silicon and its own features. The methods that turn raw ranges into a product, meaning multilateration solvers, anchor calibration, motion models, and the fusion of RF ranging with camera tracking, are patented separately, and many of the foundational filings predate the consumer UWB wave entirely. A startup shipping on iPhone hardware still needs freedom to operate at the algorithm layer. This is where licensing beats reinventing. Position Imaging licenses hundreds of granted patents across radio-frequency ranging, real-time positioning, and vision-plus-RF fusion, a portfolio cited by Apple and Bosch in their own filings. Granted patents in the pool, including US 11,774,249, US 12,079,006, US 12,066,561, and US 12,000,947, cover this method layer above the chip. Licensing turns a multi-year algorithm buildout and an open FTO question into months of integration. License the layer Apple does not sell.