White paperGNSS performance

The urban tax

The hidden cost of GNSS failure — and why a trillion-dollar navigation system fails in the places that matter most.

6 sections~8 min readFocalPoint Positioning
In brief

GNSS was engineered for open sky. The economy moved into dense cities — where signals are blocked, reflected and distorted, and accuracy collapses from a few metres to 50–100. The urban tax is the gap between the value GNSS promises and the value it delivers where its users increasingly live — paid in failed deliveries, cancelled rides, constrained autonomy, and, at the sharpest end, in minutes that translate to lives.

The failure is systematic, not random — the same street produces the same error on every pass. The techniques to close the gap are proven; they simply are not embedded in the receiver yet.

$1.4T

US value GPS has created — concentrated in cities

50–100 m

Typical urban-canyon positioning error

68%

of humanity living in cities by 2050

3–4×

accuracy recoverable in software today

Where the tax is paid

01 — The blind spot

The infrastructure that built the modern economy

Global Navigation Satellite Systems are arguably the most consequential piece of civilian infrastructure ever deployed. Yet they were engineered on an assumption that no longer reflects how the world is organised: that receivers would operate under open sky.

The global economy has spent four decades doing the opposite — concentrating into dense urban cores where satellite signals are blocked, reflected and distorted. The result is a hidden levy paid by every operator whose business depends on knowing, with confidence, where a vehicle, person or parcel actually is.

We call this the urban tax: the cumulative commercial, operational and human cost of deploying a positioning system designed for empty skies in environments defined by glass, steel and density.

$1.4T

GPS economic benefit to the US private sector since the 1980s.

NIST / RTI International, 2019

£13.62bn

Annual UK economic benefit of GNSS; a 24-hour outage would cost £1.42bn.

London Economics, 2023

€580bn

Global GNSS downstream market by 2033 — up from €260bn in 2023.

EUSPA Market Report, 2024

02 — The urban concentration of risk

The world urbanised faster than GNSS adapted

The activities that generate the largest share of GNSS-attributed value — ride-hailing, delivery, on-demand mobility, emergency dispatch, automated driving — are overwhelmingly urban.

Economic exposure to GNSS performance is not spread evenly across the Earth. It is stacked, almost perfectly, on top of the environments where GNSS performs worst.

+2.5bn

new urban residents by 2050

43

megacities over 10 million by 2030

Share of world population living in urban areas
1950 – 2050 · UN World Urbanization Prospects
70%50%35%195020002018205068%

Nearly 90% of that growth is concentrated in Asia and Africa.

03 — The urban canyon

What goes wrong between the buildings

In open sky, a standard smartphone chipset positions to within a few metres. In a dense urban canyon, that accuracy collapses. Signals are blocked by buildings and reflected off glass façades, reaching the receiver as non-line-of-sight paths it cannot distinguish from the real thing.

These errors are not random noise. They are systematic, spatially correlated and reproducible — the same narrow street between the same two buildings produces the same failure, on every pass, every day.

Because the failure is structural, the tax is not a rare event to hedge against. It is a fixed cost baked into every urban route.

Horizontal positioning error
Consumer chipset · open sky vs dense urban canyon
Open sky≈ 3 m
Urban canyon50 – 100+ m

Uber's engineering team reports standard GPS errors in cities that exceed 50 m — and sometimes 100 m — enough to place a vehicle on the wrong road entirely.

04 — Where the tax is paid

Four sectors, one recurring bill

4.1 · Automotive OEMs

The autonomy ceiling

Assisted and automated driving must know not just which road a vehicle is on, but which lane. At SAE Level 3 and above, liability shifts from driver to manufacturer — making positioning reliability a direct balance-sheet exposure.

Every kilometre of city that cannot be added to an operational design domain is revenue foregone, and a competitive gap widened.

Required lane-level accuracy for L3+
0.1–0.3 m

Roughly two orders of magnitude tighter than standard urban GNSS can guarantee today.

IEEE T-ITS · SAE localisation requirements

95%+

of ride-hailing trips occur in urban environments — so every metre of error becomes a measurable drag on lifetime value.

Cancelled trips & mis-routed drivers
Inaccurate ETAs and inflated fares
"Wasted supply" — drivers burning time
4.2 · Ride-hailing

Where every metre costs trust

Ride-hailing is a trust business. A rider dropped at the wrong corner rates lower, tips less and uses the product less — the "wrong side of the street" problem.

The tax is real; it is simply paid by the platform rather than made visible to the rider — the reason Uber acquired ShadowMaps and keeps investing in sensor fusion and 3D building models.

4.3 · Last-mile logistics

The compounding penalty

A failed first-attempt delivery is the most expensive event a logistics operator can absorb. Misdirected drivers lose minutes on every stop — minutes that compound across a route, cutting stops-per-hour and eroding thin margins.

Gig models are most exposed: non-professional drivers navigating unfamiliar geographies with consumer-grade phones — exactly where multipath errors bite.

41–53%

of total shipping cost is now the last mile

$17

average cost of a single failed delivery

+78%

projected rise in urban last-mile volume across the top 100 cities by 2030 (WEF)

Cardiac-arrest survival vs ambulance response time
≈ 5% lower survival per additional minute
90%65%50%1 min5 min10 min
4.4 · Emergency services

When seconds become casualties

The starkest version of the tax is paid in lives. The UK attributes 43% of GNSS's total economic benefit to emergency services — the single largest sector by value.

In 2024 the FCC adopted location-based routing of wireless 911 calls, estimating it could save 10,000–14,000 lives a year by cutting about one minute from response times. A system that systematically degrades in cities is not a nuisance — it is a public-health liability.

05 — The shape of the tax

The delta between value promised and value delivered

Quantifying the urban tax precisely is hard — much of it is absorbed as "cost of doing business" by operators who have already engineered around the problem with private sensor-fusion stacks and bespoke corrections. But the order of magnitude is not hard to reason about.

01

Over 80% of the UK's £13.62bn GNSS benefit sits in road transport and emergency services — both overwhelmingly urban.

02

EUSPA's €580bn 2033 market is dominated by mass-market segments whose operating environments are cities.

03

Last-mile logistics — up to 53% of shipping cost — is the segment most exposed to urban positioning failure, growing ~9% a year.

04

At L3+, OEMs assume legal liability for errors they cannot guarantee in urban canyons — gating the most valuable autonomous use cases.

The urban tax is the gap between the value GNSS creates in principle and the value it delivers where its users increasingly are. It is paid in failed deliveries, cancelled rides, constrained autonomy — and, at the sharpest end, in minutes that translate to lives.

06 — A fix is available

The fix already runs in the receiver

The urban tax is not inevitable. Software rejection of non-line-of-sight signals lets the receiver identify the reflections that corrupt an urban fix and discard them — recovering an accurate position from the very signals that mislead a standard chipset.

It runs inside the chipsets and devices already shipping — no new hardware, no new satellites. Field trials in Tokyo, Frankfurt, San Francisco and Seoul show 3–4× better urban accuracy from the software alone.

Accuracy — metre-level positioning where standard GNSS drifts by 50–100 m.
Reliability — a dependable fix on the same streets that fail today, on every pass.
Security — rejects spoofed and manipulated signals, hardening the fix against attack.
Urban positioning error
Standard GNSS vs software-enhanced (field trials)
Standard urban GNSSbaseline
Software-enhanced3–4× better

The task is not to build a new GNSS. It is to fix the one we have — precisely where it fails.

What the fix unlocks

Automotive OEMs

Lane-level accuracy in the dense cities where autonomy pays — turning a liability into operating domain the roadmap can safely add.

Ride-hailing

Right-corner pickups and accurate ETAs restore rider trust and reclaim the wasted supply lost to wrong-side-of-the-street errors.

Last-mile delivery

Correct first-attempt addresses lift stops-per-hour and cut the ~$17 failed deliveries that erode already-thin margins.

Emergency services

Faster, correctly located dispatch turns the minutes saved directly into lives saved — the sector's single largest source of value.