WeRide targets Hong Kong, Singapore for robotaxis

WeRide, a Guangzhou-based autonomous driving firm, said it plans to launch robotaxi services in Hong Kong this year and start rides in Singapore on April 1 through Grab’s app.

The company said its Hong Kong rollout would include self-driving taxis and buses, but it did not name routes or a start date. Singapore rides will initially have a safety operator to meet local rules.

Hong Kong has also approved road tests for Baidu’s Apollo Go, which said in August its fleet had logged 20,000 km in the city, and Pony.ai said it is in talks to expand there.

WeRide reported 2025 revenue of 685 million yuan (US$99.3 million) and a 1.7 billion yuan (US$246.5 million) loss, while robotaxi revenue rose to 148 million yuan (US$21.5 million).

🔗 Source: South China Morning Post

🧠 Food for thought

Implications, context, and why it matters.

Hong Kong and Singapore operate as structured markets

New deployments run under strict regulators, with clear limits for operators.
Hong Kong put a formal legal framework for autonomous vehicles in place on March 1, 2024. It replaced an ad hoc permit system under a smart city strategy that started in 2017 1.
Baidu runs approved trials under that system. Its pilot license was updated in February 2026 to allow cross-district routes up to 43.2 kilometres 2.
In Singapore, the Land Transport Authority (LTA) serves as the national transport regulator. It picked Grab to run WeRide autonomous vehicles on two set routes in Punggol. The vehicles also cleared the “Milestone 1” safety assessment for public roads 3.

Robotaxi competition shifts toward profitability

WeRide’s expansion comes as heavy losses push autonomous-driving firms to prove durable business models 4.
Pony.ai reports unit economics breakeven in Guangzhou and Shenzhen. Revenue per vehicle covers direct costs like depreciation, electricity, and maintenance 5.
In Shenzhen, Pony.ai said its robotaxis earn a one-month daily average net revenue of RMB338 (about $47) from a daily average of 23 orders per vehicle 5.
Lower hardware spending helps. Pony.ai cut the cost of its autonomous driving kit by 70% versus the prior generation 6.
Some companies use “asset-light” approaches, licensing self-driving software to partners that own and run fleets. Pony.ai uses this model 6.