Illustration: AI-generated · Inside China AI
Apollo Go: Baidu’s Robotaxi Veteran Is Testing in London — and Aiming at Germany
Since July 28, dozens of purpose-built Chinese robotaxis have been driving through the London borough of Brent — with safety operators aboard, and with two American ride-hailing giants competing to put passengers in them. Baidu’s Apollo Go, the veteran of China’s robotaxi industry, has begun UK road testing in partnership with both Uber and Lyft, with public service targeted for 2027 pending Transport for London approval. And the plan is bigger than Britain: Lyft — which bought the German taxi app Freenow for $197 million precisely for this purpose — intends to deploy Apollo Go vehicles across European markets, with the UK and Germany first in line. The service will carry a familiar name for millions of Europeans: Freenow by Lyft.
The veteran’s résumé
Apollo Go is the oldest large-scale robotaxi operation in China — Baidu has been at autonomous driving since 2013, longer than Waymo has offered public rides. Its cumulative ride count has passed 23 million, and its fleets have driven more than 350 million autonomous kilometres — over 240 million of them fully driverless — across a global footprint of 28 cities, per Baidu’s Q2 2026 disclosure. In Wuhan it operates one of the world’s largest driverless fleets in one of the world’s most chaotic traffic environments, and its international map now spans South Korea, Hong Kong, Dubai, Kazakhstan — and, alongside London, Switzerland. The London vehicles are RT6s — Baidu’s sixth-generation, purpose-built robotaxi, designed without the assumption of a driver and manufactured at a unit cost that stunned the industry when it was announced at under $30,000. Cheap, purpose-built hardware plus a decade of operating data: that is the offer Baidu brings to Europe.
The double-platform gambit
What makes the Apollo Go entry unusual is its distribution strategy: not one exclusive partner but both Uber and Lyft simultaneously. For Baidu, the platforms are interchangeable demand pipes; the asset that matters is the driver. For the platforms, the calculus is defensive — neither can afford to let the other own autonomous supply in Europe. The result is a small irony of the robotaxi age: two American platforms fighting each other to distribute Chinese autonomy on European streets, while Europe’s own automakers, as we detailed in our carmakers analysis, have no comparable robotaxi stack to offer them.
Germany is the real prize
London gets the headlines, but the Freenow acquisition points at the strategic target: Germany, Freenow’s home market, where the app is installed on millions of phones and taxi fleets already use it daily. A German Apollo Go launch would put Chinese autonomous vehicles into Europe’s largest economy through a brand Germans already trust — the softest possible market entry. It would also set up a direct confrontation with Momenta’s German L4 approval and, eventually, with whatever European alternative emerges. The 2027 timeline depends on regulators; the infrastructure — vehicles, platform, brand — is already in place.
Why it matters for Europe
Apollo Go completes the pattern our robotaxi coverage has traced: Pony.ai via Uber, WeRide via six local partnerships, and now Baidu via both American platforms at once. Three different Chinese players, three different entry models, one direction. For European regulators, the London trial is the one to watch: TfL’s decisions will become the de-facto template for how Chinese autonomy gets licensed in Europe. For the German market, the message is more immediate — the robotaxi that arrives via the familiar Freenow app in 2027 will most likely be Chinese, built for $30,000, and backed by 23 million rides of operating data. European transport policy has perhaps eighteen months to decide whether that is a procurement opportunity, a competition problem, or both.
Correction, 26 August 2026: An earlier version of this article stated 11 million cumulative rides and operations in “more than a dozen Chinese cities”. Following a note from Baidu’s international communications team, these figures have been updated to reflect the company’s Q2 2026 disclosure: 23 million cumulative rides and a global footprint of 28 cities, with more than 350 million autonomous kilometres driven. We thank Baidu for the correction.
No responses yet