Robin Li, co-founder and CEO of Baidu, has spent much of the past decade pursuing a bet that struck many observers as premature for years: that autonomous driving, not search advertising, would eventually be Baidu’s most important business. That bet reached a new milestone on July 28, 2026, when Apollo Go — Baidu’s autonomous ride-hailing platform — began real-world robotaxi testing on the streets of London through partnerships with both Uber and Lyft, marking the service’s entry into one of the world’s most scrutinized and heavily regulated markets for self-driving vehicles.
The London pilots follow multi-year global partnership agreements Baidu struck with Uber in July 2025 and with Lyft in August 2025 — deals that let Apollo Go vehicles operate directly through those companies’ existing ride-hailing apps rather than requiring Baidu to build its own rider-facing brand and user base from scratch in each new market. Lyft’s initial London fleet numbers in the dozens of vehicles, with plans to scale into the hundreds pending regulatory approval; Uber’s first UK pilot is expected to begin in the first half of 2026. By mid-2026, Apollo Go’s global footprint had grown to roughly 22 cities and more than 250,000 weekly trips, with the service also active in South Korea — its first expansion into an Asian market beyond mainland China and Hong Kong — and in the Middle East through partnerships in Dubai and Abu Dhabi.
Domestic Scale Backing the Global Push
Apollo Go’s international expansion is underpinned by substantial domestic scale within China, where the service operates in around 20 cities and posted a weekly order peak exceeding 300,000 trips in the fourth quarter of 2025, with cumulative global rides surpassing 20 million by February 2026. Li has framed the next phase of the business less around simply adding more vehicles and cities, and more around integrating robotaxis into how cities actually function — describing the current stage as centered on “how robotaxi services fit more naturally into public transportation, city operations and everyday life,” rather than treating autonomous vehicles as a novelty layered on top of existing transit systems.
A Broader AI Strategy Beyond Robotaxis
Li’s robotaxi bet sits alongside a parallel push into large language models and generative AI at Baidu, the company long known as China’s dominant search engine. Baidu unveiled ERNIE 5.0, a natively multimodal foundation model capable of jointly processing text, images, audio, and video, at its Baidu World 2025 event, and reported its ERNIE AI assistant had reached 202 million monthly active users by December 2025. In February 2026, Baidu also announced its first-ever dividend and a $5 billion share repurchase program, a shareholder-return signal that came alongside continued heavy capital spending on AI infrastructure, including two new custom AI chips — the Kunlunxin M100 and M300 — slated for release in 2026 and 2027, aimed at reducing Baidu’s dependence on foreign-made AI chips amid continued U.S. export restrictions.
Years of Being Called Premature
Li has publicly described AI as a decade-long conviction that China’s broader technology industry has only recently caught up to, having pushed Baidu into autonomous driving research and large-scale AI investment well before the current generative-AI boom made such spending fashionable across the industry. That early, sustained investment left Baidu with a head start in the specific, capital-intensive infrastructure autonomous driving requires — mapping, sensor fusion, and years of accumulated real-world driving data — even as the company faced periods of investor skepticism about whether the payoff would ever materialize, particularly as Baidu’s core search advertising business faced increasing competition from short-video and AI-native alternatives.
Competing Against a Crowded Global Field
Apollo Go’s international push places it in direct competition with Alphabet’s Waymo, which has established an early lead in the U.S. robotaxi market, as well as with other Chinese autonomous driving companies such as WeRide, which have also pursued international expansion strategies of their own. Li has positioned Apollo Go’s Uber and Lyft partnerships as a structural advantage over rivals building standalone consumer apps market by market, arguing that plugging into ride-hailing platforms with existing, established user bases in each city allows faster real-world deployment than trying to win rider trust and adoption from zero in every new market Apollo Go enters.
Baidu founded Apollo, its broader autonomous driving program, in 2017, and has spent the years since building out both the underlying self-driving technology and the regulatory relationships in individual Chinese cities needed to operate driverless vehicles commercially — groundwork that has since translated into faster regulatory conversations in some international markets, given Baidu’s now-established track record operating safely at meaningful scale within China.
What This Means for Philippine Founders
Robotaxi and autonomous vehicle technology remain distant from near-term deployment in Philippine cities given the country’s traffic infrastructure, regulatory environment, and road conditions, but Baidu’s platform-partnership model — plugging autonomous vehicle capability directly into an existing ride-hailing app’s user base rather than building a separate consumer brand — is directly applicable today to Philippine mobility and logistics startups considering how to scale a new transportation technology without also having to win riders’ trust and habit formation from scratch. Filipino founders in adjacent fields like AI-powered logistics routing, fleet management, or last-mile delivery should also watch how Baidu’s ERNIE-based AI stack and custom Kunlunxin chips develop, since a cost-competitive Chinese AI infrastructure stack could eventually offer Southeast Asian startups a lower-cost alternative to U.S.-based cloud AI providers.
Share this article