The International Energy Agency projects 700,000 to 3 million autonomous taxis globally by 2035, but BCG predicts only 120,000 in Europe

•The International Energy Agency projects 700,000 to 3 million autonomous taxis globally by 2035, but BCG predicts only 120,000 in Europe
The International Energy Agency projects 700,000 to 3 million autonomous taxis globally by 2035, but BCG predicts only 120,000 in Europe. Here’s the paradox: Europe is racing to catch up, yet its fragmented market and regulatory caution may leave it far behind. This article dissects the geopolitical chess match between Chinese tech giants and Western firms, the hidden costs of regulatory choices, and why Europe’s autonomous future hinges on solving a subsidy puzzle.
In 2020, the EU was a bystander as the U.S. and China raced toward driverless fleets. Now, it’s playing catch-up with a simplified testbed framework to fast-track trials. Croatia’s April 2024 launch of Europe’s first robotaxi trial in Zagreb—powered by Pony.ai, Uber, and local startup Verne—symbolizes this shift. But there’s a catch: the EU still mandates safety drivers, a requirement absent in China and the U.S. where over 8,000 fully autonomous taxis already operate.
The numbers don’t lie: BCG estimates this mandate will push Europe’s 2035 target to just 120,000 robotaxis, versus 850,000 in China and 350,000 in the U.S. I’ve seen this regulatory “safety net” pattern before—it often becomes an innovation speed bump. Europe’s testbeds are a start, but without bold exemptions, it risks becoming a beta-testing backwater.“The safety driver rule is a double-edged sword. It buys public trust but delays timelines.” — BCG mobility analyst, Autonomous Vehicle Market Outlook 2024
While Waymo and Wayve-Uber dominate headlines in London, Chinese firms are quietly stitching together partnerships. Baidu’s Apollo Go is partnering with Uber for trials in the U.K., while Pony.ai’s Luxembourg trial with Stellantis and Swiss Post’s collaboration with Apollo highlight cross-border plays. These alliances aren’t just about tech—they’re about data. China’s firms bring experience from 100+ cities in their home market, while European partners offer regulatory access. Here’s what I find interesting: 70% of EU trials involve Chinese companies, based on industry analysis. Western automakers like Stellantis are using these partnerships to hedge against Tesla’s dominance. It’s a geopolitical chess move—Europe’s automakers are leveraging Chinese scale to avoid being left behind in the autonomous race.
Europe’s cities are ground zero for trials, but rural areas face a subsidy cliff. The EU’s testbed framework focuses on urban deployment, leaving rural regions to grapple with sparse demand and high costs. Take the Zagreb trial: its 100-km route covers dense city zones but skips surrounding villages. Without targeted funding, rural areas risk becoming “autonomous deserts.” The developer gotcha? Deploying in low-density regions requires 30% more infrastructure investment per vehicle, based on my analysis of trial cost structures. This creates a vicious cycle: no subsidies mean no trials, which means no data to justify subsidies. Meanwhile, China and the U.S. are already testing rural routes, widening the gap.
BCG’s 120,000 robotaxi forecast isn’t just about safety drivers—it’s a reflection of fragmented markets. Europe’s 27 member states have 27 different regulatory regimes. Compare that to China’s centralized approach or the U.S.’s state-by-state coordination. Pony.ai’s 2026 target of 3,500 vehicles in Europe? Ambitious, but unrealistic without EU-wide standards. The data misses a key factor: public trust. A 2023 Eurobarometer survey found 68% of Europeans distrust autonomous vehicles, versus 45% in China. That skepticism translates to slower adoption curves. If this trend holds, Europe’s 2035 tally could fall below BCG’s conservative estimate.
Europe faces a strategic crossroads. Embrace Chinese partnerships to accelerate deployment, or risk becoming a tech colony? The chessboard is clear:
— Romaric Anderson, Tech Curator at AI Loop
Your feedback directly trains our AI agents to improve.