UWB vs BLE AoA for Pallet Tracking: Real Costs at 500k Sq Ft

Why does UWB hold 10 to 30 cm where BLE AoA drifts to meters?

Start with the physics. An 802.15.4z UWB pulse spreads 499.2 MHz of bandwidth, so a receiver can pick out the first-arriving signal to well under a nanosecond. Light travels about 30 cm per nanosecond, which is why field deployments hold 10 to 30 cm even between loaded racks. Reflections off steel arrive late and get gated out of the time-of-flight measurement.

BLE AoA takes the opposite approach. A locator with an antenna array measures phase differences on a narrowband 2.4 GHz carrier and computes an angle to the tag. Vendor datasheets quote 1 to 2 degrees, measured in open air. Between full rack rows, multipath merges with the direct signal and angle error grows to 5 degrees or worse. At 15 m of range, 5 degrees is 1.3 m of position error. Ten degrees is 2.6 m.

Standard rack bays place pallet positions about 1.3 m apart center to center. UWB resolves the position. BLE AoA resolves the bay. Steel racking decides this comparison.

What does infrastructure cost across 500,000 square feet?

UWB TDoA needs each tag visible to four or more anchors, and racking blocks line of sight, so plan one anchor per 2,500 to 4,000 sq ft. That is 125 to 200 anchors at $400 to $800 each, plus a clock synchronization layer, wired or over-the-air. BLE AoA locators run cheaper per unit, $300 to $600, but angle-only positioning needs clean geometry down every aisle, so counts land at 150 to 250 units for the same footprint.

Installation is where quotes diverge from invoices. Every mount at 30 to 40 ft requires a boom lift, and a PoE home run in conduit typically prices at $400 to $900 per drop. Two hundred drops at $650 is $130,000 before a single radio powers on.

With commissioning, expect UWB at $250,000 to $450,000 all-in for 500,000 sq ft, about $0.50 to $0.90 per square foot. BLE AoA lands at $150,000 to $280,000, about $0.30 to $0.55. Cabling costs more than the radios.

How do tag economics change the math at 25,000 pallets?

A 500,000 sq ft facility commonly carries 20,000 to 30,000 pallets, so price tags at 25,000 units.

• BLE tags: $5 to $12 each, coin cell, 3 to 5 years at a slow blink rate. Fleet cost $125,000 to $300,000.
• UWB tags: $15 to $35 in 2026 volume pricing, down sharply since 802.15.4z silicon reached phone-scale production volumes. Fleet cost $375,000 to $875,000.

That flips the totals from the infrastructure section. All-in, BLE AoA lands around $300,000 to $600,000 and UWB around $650,000 to $1.3 million. Two levers pull UWB back down. Dropping the blink interval to once per 10 seconds stretches tag batteries past three years and cuts network load, and reusable license-plate tags that move between pallets can cut fleet size 30 to 50 percent in facilities with fast turns. Run the tag math before the anchor math. At pallet scale, tags outweigh anchors.

When is BLE AoA still the right choice?

Bay-level accuracy answers real questions, and if your workflow only asks those questions, buying centimeters wastes money.

• Cycle counting to the bay. If a picker can scan one of three candidate pallets, 1.5 m accuracy is enough.
• Dock door in and out. Detecting which of two doors a pallet crossed needs 2 to 3 m of separation, well within AoA range.
• Staging lane occupancy. Lanes 2.5 m wide resolve cleanly at BLE AoA accuracy.

The pattern that works in 2026 is hybrid: BLE AoA across the full footprint for presence and zone data, UWB only in shipping verification and staging areas where loading the wrong pallet triggers a retailer chargeback. Dual-radio tags that speak both protocols ship from multiple vendors now, so one tag fleet serves both layers. Match the radio to the workflow, per zone.

Which line items blow past the original quote?

The hardware quote is the floor, not the number. Budget for what integrators price after the contract is signed.

• Lift time and after-hours labor. Installs in a live facility happen nights and weekends at premium rates, with boom lift rental on top.
• Network capacity. One hundred fifty to 250 PoE devices need switch ports, and a 500,000 sq ft run usually needs fiber backhaul between IDF closets.
• Calibration, twice. Initial RF survey and anchor calibration, then recalibration after every major re-slotting, because new inventory geometry changes multipath.
• Software fees. Per-tag pricing of $0.10 to $0.50 per month reads as small print until 25,000 tags makes it $30,000 to $150,000 per year, forever.
• WMS integration. Mapping location events to your slotting model is weeks of work, not a connector toggle.

A working rule: these items add 40 to 100 percent on top of hardware. Budget the lift time before the hardware.

Should you build the positioning stack or license it?

The gap between a datasheet number and a working deployment is not the radio chip. NXP and Qorvo sell UWB silicon to anyone. The gap is the algorithm layer: first-path detection in multipath, anchor time synchronization, fusing RF ranging with motion data so a stationary pallet does not wander on the map. Teams that build this from scratch spend 18 to 24 months rediscovering solved problems, and much of that ground is already patented, which puts a freedom-to-operate question under the whole project.

Position Imaging licenses hundreds of granted patents in real-time positioning, radio-frequency ranging, and computer vision, including US 11,774,249, US 12,079,006, US 12,066,561, and US 12,000,947, from a portfolio cited by Apple and Bosch in their own filings. Licensing proven IP means your engineers spend their months on the warehouse product, not on re-deriving ranging math that already exists with a patent number on it. License the hard part, ship the product.