Last week, Nevada regulators handed Tesla permits for up to 5,000 robotaxis in the Las Vegas area, with Waymo and Uber each cleared for another 1,000. That is roughly 7,000 permitted autonomous vehicles for a single US metro area, in a single announcement.
Meanwhile, in Punggol, Singapore, Southeast Asia’s most advanced public robotaxi trial, Grab and WeRide are running 11 vehicles along two fixed routes, free of charge, with commercial fares still pending.
Also Read: Grab makes strategic bet on WeRide to drive autonomous mobility in SEA
The contrast is stark enough to look like a failure of nerve. It isn’t. But it is a warning that Southeast Asia’s autonomous vehicle strategy needs to become a lot more deliberate before the gap becomes a gulf.
The scale gap is real, and it’s not just about money
Start with what Las Vegas actually signals. Nevada’s willingness to license fleets in the thousands, rather than tens, marks a shift from “pilot” to “infrastructure.” Tesla, Waymo and Uber are now treating a single city as a live commercial market, not a proof of concept. The US is betting that regulatory boldness, not just capital, is the scarce resource in the robotaxi race.
Southeast Asia has capital. Grab is Southeast Asia’s largest ride-hailing and delivery operator, and it has spent the past two years building exactly the kind of partnerships this moment calls for: an investment in Chinese autonomous driving firm WeRide and a separate tie-up with Michigan-based May Mobility aimed at adapting self-driving systems to the region’s roads. What the region has not had, until now, is a Nevada-style regulator willing to license fleets at four-digit scale.
That caution is not irrational. It is the product of genuinely harder conditions.
Why Southeast Asia moved slower and why that’s defensible
Singapore’s own roadmap targets only 100 to 150 self-driving vehicles by the end of 2026, a rounding error next to Las Vegas’s new permits. But Singapore’s roads, like most of the region’s, mix motorcycles, informal transport, unpredictable pedestrian crossings and left-hand traffic patterns that US autonomy stacks were never trained on. May Mobility’s own framing of the challenge is instructive: its CEO has said the plan is to bring the company’s autonomy system to the region as early as regulators allow, without committing to a specific market first. That is an admission that the technology, not just the paperwork, still needs local adaptation.
Also Read: Can autonomous delivery vehicles handle the chaos of real roads?
The caution is also informed by recent failures elsewhere. Robotaxi passengers have been left stranded for hours when a fleet’s software or connectivity failed, and a self-driving vehicle in China reportedly ended up in a construction pit. A regional operator scaling to thousands of vehicles before the technology has proven itself on SEA’s specific road conditions would be inviting exactly that kind of incident, at a much larger, more damaging scale.
So the 11-vehicle fleet in Punggol isn’t timidity. It’s a deliberate, government-coordinated test run, with Grab’s driver-partners retrained as safety and remote operators rather than displaced outright. That is a meaningfully different model from Nevada’s regulatory greenlight-and-scale approach, and arguably a more exportable one, for markets that cannot afford Las Vegas-style mistakes.
The leapfrog Southeast Asia can still make
Here is where the region has a genuine opening, rather than just an excuse. Grab is not simply importing American or Chinese autonomy technology; it is feeding its own mapping and routing data into May Mobility’s system specifically so the technology learns Southeast Asian traffic before it scales.
That is the leapfrog move: skip the “American roads first” assumption entirely, and build an autonomy stack whose first real-world competence is in the traffic conditions most of the world’s fast-growing cities actually have, not the wide, well-marked boulevards of Las Vegas.
If Southeast Asia gets this right, the region doesn’t just catch up to Nevada’s numbers eventually. It ends up holding the more commercially valuable asset: autonomous driving systems proven on the chaotic, mixed-mode traffic that characterises most of Asia, Africa and Latin America, rather than systems calibrated for wide American arterial roads. Nevada is optimising for scale in a forgiving environment. Singapore, if it moves deliberately, is optimising for robustness in an unforgiving one and robustness travels further.
What has to happen next
Three things need to move faster than they currently are.
First, regulators across the region, not just Singapore’s Steering Committee on Autonomous Vehicles, need clearer, published pathways from pilot to commercial fare, so operators can plan capital deployment instead of guessing at timelines.
Second, insurance and liability frameworks for mixed autonomous-human traffic need to exist before fleets scale past a few dozen vehicles, not after an incident forces the issue.
Third, the labour transition Grab has started — retraining driver-partners as safety and remote operators — needs to become an explicit regional policy commitment, not a single company’s goodwill gesture, given how many SEA livelihoods depend on ride-hailing and delivery work.
Also Read: Autonomy vs anarchy: How do we secure the future of autonomous transportation?
None of this means Southeast Asia should try to match Las Vegas vehicle-for-vehicle. It shouldn’t, and it can’t — not yet, and possibly not for years. But the region does need to stop treating its caution as a plan in itself. Caution bought Singapore a working 11-vehicle trial with real ridership data and a retrained workforce. It has not yet bought the region a credible answer to the question Nevada just asked out loud: what happens when robotaxis stop being a pilot and start being a market?
Southeast Asia has the ingredients — the superapp distribution, the local road data, the capital, the regulatory relationships — to answer that question on its own terms rather than importing someone else’s answer wholesale. What it doesn’t have yet is a timeline. Until it does, Las Vegas gets to write the scale story, and the region only gets to write the caveat.
The post Southeast Asia isn’t losing the robotaxi race. It’s running a different one appeared first on e27.