E-scooters

Are e-scooters safe?

They are about as risky as cycling per mile, but nobody measures it precisely. No government publishes deaths per mile for e-scooters, so we put the figure together from two sources. Germany's police recorded 21 e-scooter riders killed in 2023, and the national travel survey for that year, Mobilität in Deutschland, found e-scooters used for about 1% of the country's 250 million daily trips. That is about 0.02 to 0.05 micromortsMicromortA one-in-a-million chance of death. It lets you compare the risk of very different activities on one scale.More in the glossary per ride, or roughly 1 to 4 micromorts per 100 miles, with about 2 the middle estimate. For comparison, a British cyclist faces about 2.3 per 100 miles and a car occupant far less.

Most rides are short, so the risk of any one ride is small. The bigger problem is injury. E-scooters put riders on small wheels at bicycle speeds, usually without a helmet, and the injuries that follow are often to the head.

How risky is an e-scooter compared with a bike or a car?

Mode	Deaths per billion km	MicromortsMicromortA one-in-a-million chance of death. It lets you compare the risk of very different activities on one scale.More in the glossary per 100 miles	Source
Car occupant (Britain, 2024)	1.0	0.16	DfT
Bicycle and pedelec (Germany, 2023)	about 10	about 1.6	Destatis, MiD 2023
E-scooter (Germany, 2023)	about 8 to 23	about 1.2 to 3.7	Destatis, MiD 2023
Rental e-scooter (Europe, 2021)	15	2.4	Micro-Mobility for Europe

How we got the German rows: the survey rounds the e-scooter share to 1% of trips, which could mean anything from 0.5% to 1.5%. Dividing the 21 deaths by that many trips a year gives 15 to 46 deaths per billion trips. MiD says rental rides average just over 2 km, so we assumed 2 km a ride to convert this to distance. Longer private rides would lower the rate. The cycling row uses the same survey and the 446 cyclists killed in 2023. These are rough figures. The survey's own e-scooter share of kilometres travelled is too coarse to use.

The rental row comes from Micro-Mobility for Europe, a group of operators including Bird, Bolt, Dott, Lime, Tier and Voi, which reported 7 rider deaths over 461 million km. It has an interest in the result, but it agrees with the German estimate. Britain recorded 10 e-scooter riders killed in 2025 and the US at least 38 in 2024, but neither country counts e-scooter miles.

Deaths are rising with use. German police counted 33 riders killed in 2025, up from 21 in 2023, and e-scooter crashes with injuries rose 38% in 2025 alone. Half of all injured riders were under 25.

Why the injuries are worse than they look

The US Consumer Product Safety Commission estimates that emergency departments treated about 380,000 e-scooter injuries between 2017 and 2024, and 79,300 in 2024 alone. In its follow-up of 2024 cases, only 18% of injured riders were wearing a helmet, and a third were carrying something while riding. In Norway, about one shared-scooter rider in ten reported an accident needing a doctor in the past year, or 48 per million km.

Head injuries stand out. A 2026 study of the national trauma registry for England and Wales found e-scooter riders had 3.45 times motorcyclists' risk of traumatic brain injury among trauma patients, and 1.74 times cyclists'. Brain injury made up more than a third of adult e-scooter trauma. Small front wheels can stop dead in a pothole or against a kerb and throw the rider forward.

Other people get hurt too. In Britain in 2025, 235 pedestrians were injured in collisions with e-scooters, 54 of them seriously. CPSC knows of 6 US pedestrians killed by e-scooter riders in 2017–2024.

What kills e-scooter riders

Cars, mostly. Of 206 US e-scooter deaths known to CPSC in 2017–2024, 140 involved a motor vehicle. Loss of control into kerbs or fixed objects caused 20 more, and unexplained falls 21. Battery fires caused another 15 deaths. In Germany, though, 16 of the 33 riders killed in 2025 died in crashes with no one else involved.

Speed, alcohol and night riding matter a great deal. German police blamed alcohol in 10.9% of the faults they recorded for e-scooter riders in 2025, against 7.7% for cyclists. When Helsinki capped rental scooters at 20 km/h and banned them late on Friday and Saturday nights, hospital-treated injuries fell from 19 to 9 per 100,000 rides, a 50% drop after allowing for weather.

How to ride an e-scooter more safely

1. Wear a helmet. A bicycle helmet is the minimum.
2. Keep the speed down, especially on poor surfaces. Many rental scooters cap at 20 to 25 km/h; private ones can go much faster.
3. Do not ride after drinking. Night and alcohol drove much of Helsinki's injury load.
4. Look for potholes, tram tracks and kerbs. Small wheels cannot ride through what a bike can.
5. Keep both hands on the bars. Do not carry bags or phones in your hand.
6. Ride alone. German police counted 822 passengers injured or killed in 2025, though a second rider is not allowed.
7. Use lights and be visible, and ride in bike lanes where you can.
8. Check the law. In Britain, private e-scooters are not allowed on public roads or pavements; only rental scooters in about 20 trial areas are.
9. Charge safely. Use the original charger, do not charge overnight or in a hallway that is your only way out, and replace damaged batteries.

How we scored it

Fatality risk (75) sits in the 1–5 micromortMicromortA one-in-a-million chance of death. It lets you compare the risk of very different activities on one scale.More in the glossary band per 100 miles, close to cycling. Injury risk (45) is the weak point: hospital visits are common and head injuries severe. Controllability (60) reflects how much speed, alcohol and helmets change outcomes. Rescue and environment (88) is good because nearly all riding is in cities. That gives 67 overall. A scooter is safer than a motorbike per mile, but far riskier than driving.

Our confidence is medium. The death counts are official, but the rate per mile rests on a rounded survey share and an assumed ride length, so treat it as a range rather than a precise number.