Driverless Taxis Could Come to Singapore by 2028

Waymo plans to offer taxi rides with no driver to the public in Singapore by 2028. The Verge

Vehicles will begin arriving in Singapore in the coming months to prepare for mapping and testing. Mapping, making a detailed street record the car can follow, and testing with a human safety driver on board is scheduled for 2027. It is like a learner driver with an instructor ready to take over. Fully driverless public service is the stated target for the following year, subject to regulatory clearance.

That clearance is explicit. Waymo will need approval from Singapore's Land Transport Authority before accepting passengers. The sequence is mapping, then supervised autonomous testing, then an application for passenger operations. No driverless fares before the regulator signs off.

Singapore is one of several new markets in parallel. Waymo plans to launch a robotaxi service, a taxi with no driver, in London by the end of 2026, in Tokyo in 2027, and in Munich in late 2027. Each market has its own road layouts, traffic mix and approval path, but the pattern is consistent: place vehicles early, map, test supervised, then remove the driver.

The background mileage comes from the United States. Waymo, a unit of Google parent Alphabet, has around 4,000 vehicles operating in about 15 U.S. cities. Reuters The company has offered robotaxi rides in Los Angeles, San Francisco and Phoenix.

Its published safety analysis covers more than 220 million fully autonomous miles through the end of March 2026, across five operating areas, including Atlanta for the first time. Waymo Compared with human drivers in the same areas, the Waymo Driver, the company's self-driving system, was involved in 94% fewer crashes causing serious or fatal injuries. The reduction was 82% for crashes in which an airbag deployed. It was 82% for crashes involving any reported injury.

Numbers for people outside cars are separated out. Waymo reported 93% fewer injury-causing crashes involving pedestrians than the human benchmark, the local human crash rate. The reduction was 84% for injury-causing crashes involving cyclists. It was 84% for injury-causing crashes involving motorcyclists.

Atlanta is now counted as a separate service area. The analysis there covers more than 5.4 million fully autonomous miles. In Atlanta, the Waymo Driver was involved in 94% fewer airbag-deployment crashes than the human benchmark and 86% fewer injury-involving crashes. System-wide, Waymo said it now drives more than 4 million miles every week.

The broader context here is competitive capacity, not just Waymo execution. Uber and Nvidia had planned a robotaxi rollout to 28 cities globally by 2028, covering North America, Europe, Australia and Asia. Reuters That plan frames Singapore less as an isolated pilot and more as one contested market among many where fleet operators, automakers and chip suppliers are negotiating for road time and regulatory trust.

In my view, the Singapore timeline reads as deliberate rather than conservative. Left-hand traffic, dense urban corridors, heavy rain and strict enforcement make it a demanding test. Starting with mapping vehicles months in advance, then a full year of supervised testing, gives engineers time to adjust the system for local tricky situations before carrying paying passengers. Approval is the gate. Mileage is the argument for passing through it.

Looking at what this means for riders, watch the handoff points. Supervised testing checks the driving software. Driverless operation checks the whole system: dispatch, fleet response, incident handling, data logging for the regulator. Crash-rate comparisons matter because regulators and insurers price risk on benchmarks. If the 2027 testing phase produces clean, auditable data in Singapore conditions, 2028 becomes a procurement and policy discussion. That is how robotaxi networks scale, city by city, permit by permit.