Geely’s Eva Cab Robotaxi: Built to Forget the Human Driver
Unveiled at the Beijing Auto Show, the Geely Eva Cab robotaxi is a showcase of autonomous EV technology designed from day one to operate without a human driver. Rather than adapting a consumer car, Geely built a dedicated robotaxi electric vehicle with no steering wheel or pedals, freeing up the cabin for passengers and comfort-focused features. Under the skin, a high-performance AI system with 196 billion parameters and 1,400 TOPS of computing power drives what Geely calls its “World Action Model,” continuously running “micro-deductions” to handle up to 99% of traffic scenarios, including tricky unmarked roads. The Eva Cab layers this software onto 43 sensors, including cameras and LiDAR, to achieve a claimed 4-millisecond reaction time—about three times faster than a human—while executing difficult maneuvers with a 95% success rate. Geely targets 2027 mass production with Caocao Mobility, but regulators and real-world validation will decide how quickly these robotaxis scale.
Inside the Tesla FSD Experience: Powerful, but Still Supervised
While Geely chases a fully driverless robotaxi, Tesla is iterating on autonomy through its existing fleet. In Europe, owners in the Netherlands are currently trialling Full Self-Driving (Supervised), offering a glimpse of Tesla FSD experience on public roads. Before activating the feature, drivers must watch a tutorial and pass a short in-car quiz, reinforcing that FSD remains a driver-assistance system requiring constant oversight. Early user reports describe the first commute with FSD engaged as “surreal,” with the car handling most of the driving once activated, yet owners are clearly primed to stay ready to intervene. Crucially, the “Supervised” label is not just legalese; it underscores that today’s Tesla system is an advanced co-pilot rather than a replacement for the driver. Other EU countries are still deliberating approval, highlighting how regulations lag behind software updates and how uneven the path to autonomous EV technology remains across markets.
Tesla’s Autonomy-First Roadmap and the Roadster Exception
On Tesla’s latest roadmap, Elon Musk signalled a dramatic philosophical shift: in the future lineup, the only Tesla designed for full human control is expected to be the second-generation Roadster. Everything else, from mainstream models to the upcoming Cybercab, is being positioned as autonomy-first, evolving toward self-driving via new hardware and Full Self-Driving software. Vehicles built before 2023 are constrained by older hardware, while newer Hardware 4 cars are expected to unlock higher levels of automation through updates, supported by retrofit or upgrade paths for legacy owners. The planned Cybercab concept eliminates steering wheel and pedals altogether, mirroring the cabin-first design of the Geely Eva Cab robotaxi. In contrast, the Roadster is explicitly not being engineered around safety or autonomy as primary goals, instead acting as a pure driver’s halo car. This deliberate outlier underscores Tesla’s bet that everyday mobility will be dominated by autonomous EVs, with human-driven performance cars becoming niche.
Robotaxi vs. Upgraded Consumer EVs: Two Paths to the Same Future
Geely and Tesla are converging on the self driving car future from opposite directions. The Geely Eva Cab robotaxi represents a clean-sheet approach: a purpose-built autonomous EV, packed with sensors and compute, designed for fleet deployment and operated exclusively by software. Its 4-millisecond reaction time, 43-sensor array and removal of all driving controls demonstrate a commitment to Level 4/Level 5-style use cases from day one. Tesla, by contrast, is turning consumer EVs into robotaxi electric vehicles via over-the-air software and incremental hardware upgrades, betting heavily on a camera-first perception stack and existing owner base. Purpose-built robotaxis may offer higher safety redundancy and better-optimised interiors, but they require massive upfront fleet investment and regulatory approvals city by city. Upgraded consumer EVs can scale more organically, yet must grapple with mixed modes—humans and software sharing control—and with regional regulators that may approve supervised FSD long before they accept fully driverless, control-free cabins.
Regulation, Ethics and Business Models in Southeast Asia
For markets like Malaysia and Southeast Asia, the biggest barriers to fully driverless EVs are less about chips and more about rules, roads and trust. Dense, often chaotic traffic, variable road markings and heavy two-wheeler presence will test whether systems like Geely’s World Action Model or Tesla’s camera-based FSD can truly handle 99% of real-world scenarios. Regulators must decide how to certify AI drivers, apportion liability in crashes and govern data use from sensor-rich robotaxi fleets. Ethically, questions loom around job losses in driving-focused industries and how to ensure equitable access if autonomous services cluster in affluent urban cores. Still, the potential business models are compelling: operator-owned robotaxi electric vehicle fleets for ride-hailing, autonomous logistics vans for last-mile delivery, and privately owned EVs that join networks as part-time robotaxis. In practice, Southeast Asia is likely to see a hybrid landscape where autonomous fleets and human-driven EVs coexist for many years.