A Chinese robotics startup is already beating Western rivals at scale

DHL has moved a fleet of fully electric, driverless vehicles from Chinese startup Zelostech into live operation at its Advanced Regional Center in Singapore, the first DHL site worldwide to run the technology beyond a trial. The small fleet works around the clock, hauling up to 1.5 tons per trip across roughly 40 daily runs that total 28 kilometres. DHL says the vehicles cost about half as much to run as diesel trucks.

The vendor, not the buyer, is the story. Zelostech reports about 4,000 autonomous vehicles already working across more than 300 cities, a scale most Western logistics firms have yet to approach.

Zelostech has roughly 4,000 autonomous vehicles running across more than 300 cities. Founded in China in 2021, the startup has built a deployment base that dwarfs what most Western logistics players have managed, and almost none of those Western firms have left the pilot stage.

That gap is the real news inside an otherwise routine deployment announcement. On June 6, 2026, DHL confirmed it is running Zelostech’s driverless electric vehicles in full operation at its Singapore campus, moving cargo between on-site buildings without a driver in the seat. The vehicles are not navigating public streets. They are doing the dull, repetitive work that fills the middle of every supply chain.

Scale, it turns out, is the weapon here. A Western operator validates its software across a few fixed routes and a handful of test sites. A Chinese vendor like Zelostech drops thousands of units into closed campuses and harvests operating hours by the million. The learning curve bends faster when you have that much real-world data, and that is the part Western procurement teams have not fully priced in.

A Chinese vendor is logging the hours Western rivals cannot

The technology does one narrow thing well. Each autonomous vehicle carries up to three pallets between fixed points inside DHL’s Advanced Regional Center (ARC), running 24 hours a day with no human aboard. This is the first DHL location globally to graduate a Zelostech vehicle from an innovation challenge to permanent duty. The routes are known, the movements repeat, and the safety rules are tight — exactly the conditions where today’s autonomy actually works.

Wei Kieng Eunis Hew, Managing Director of DHL Supply Chain Singapore, called it a “full-scale operation” meant to “transform the way we move goods between sites” and deliver “smarter, greener logistics solutions.” That is the business framing. The engineering framing matters more.

Closed campuses let Zelostech skip the hardest problem in autonomy: unpredictable public roads. A close Western parallel is Nuro and Gatik in North America, which run small driverless trucks on fixed short-haul routes. Both stayed largely on public roads under strict geofencing and safety drivers, slowed by tighter liability rules. Zelostech went the other way and scaled inside fenced industrial parks instead. The result is a vendor accumulating reliability data far faster than its Western counterparts.

Here is the twelve-to-eighteen-month implication. If similar vehicles get imported or copied in Europe and North America, the Chinese firms arrive with millions of logged operating hours and a cost structure no pilot can match. The deployment count is real. Whether it proves quality at scale is the question the numbers cannot answer yet.

The middle mile is where automation actually pays

Forget robotaxis. The money in autonomy is moving toward the unglamorous “middle mile” — the constant transfers of goods between warehouses, sorting centres, and hubs. These links are repetitive, congestion-prone, and chronically short on staff. They are also fenced, mapped, and rule-bound, which is precisely why a vehicle that fails on a city street can succeed on a campus.

Two structural forces are pushing the same way. Fortune Business Insights projects rising e-commerce volumes and safety pressures as the core drivers of adoption through 2034, with Asia-Pacific expected to grow fastest. McKinsey points to the other half: a labour cliff, with its European driver-shortage analysis calling automation one of several levers firms will pull to hold service levels.

So the Singapore deployment is small, but the signal is not. A Chinese vendor has done the hard work of scaling in closed environments while Western freight firms validated safety on open roads. The 4,000-vehicle fleet is not proof of better engineering. It is proof of more practice — and in autonomy, practice is the asset that compounds.

Beyond the headline

The bigger picture

The shift underway is from headline-grabbing robotaxis to the quieter, more automatable tasks inside logistics campuses. Software-defined fleets are becoming part of warehouse infrastructure itself, turning labour-intensive, low-margin links into capital-intensive, data-rich assets that can be copied site to site.

The reach

One overlooked player is the institutional investor holding logistics real estate. Landlords offering “AV-ready” parks — standard layouts, connectivity, charging — gain a pricing and occupancy edge as automation proves itself. That shapes where Western retailers base regional hubs and which cities win the next wave of e-commerce build-out.

What isn’t being said

Corporate messaging stays quiet on data ownership. These fleets generate detailed maps of warehouse layouts, cargo flows, and bottlenecks. When a foreign startup’s stack collects that telemetry, the open questions are who can access it, how it meets national cybersecurity rules, and whether governments will eventually treat such data as strategic infrastructure rather than an efficiency metric.

What the Singapore template means for your decisions

With a Chinese vendor now running live operations in Singapore and the technology poised to spread across Asia-Pacific campuses over the next 12 to 18 months, three groups face concrete choices.

• Corporate strategists tracking logistics

  Pull the executive summary of the autonomous trucks market report from Fortune Business Insights and test its Asia-Pacific growth assumptions against your own sourcing and distribution exposure. If your regional hubs depend on Chinese vendors for automation, map that dependency now, before competitors lock in supply.

• Policy and compliance teams

  Review the Land Transport Authority of Singapore’s autonomous vehicle framework at lta.gov.sg to see how regulators structure exemptions, safety conditions, and data-sharing. Use it to benchmark against AV rules in your own jurisdiction, particularly the data-governance gaps that scale will eventually force open.

• Investors in logistics real estate

  Watch for a formal DHL announcement on expanding Zelostech beyond Singapore. If it lands, it signals a validated business case worth tracking through “AV-ready” property plays. If it stalls, the technology likely stays in pilot mode while DHL assesses performance and labour impact.

Explainer

Autonomous vehicle

A vehicle that drives itself using sensors, mapping, and onboard software, with no human at the controls. In logistics, the most reliable use is inside closed campuses where routes are fixed and obstacles are predictable, rather than on open public roads. Zelostech’s units carry up to three pallets and run continuously, a duty cycle no human driver could match.

Advanced Regional Center

DHL’s term for a large, multi-building logistics campus serving a region. The Singapore ARC is the first DHL site worldwide to move a Zelostech vehicle from a contest entry to permanent operational use. Its fenced, single-tenant layout is what makes the autonomy work — a harder problem in multi-tenant parks with less central control.

Land Transport Authority

Singapore’s lead transport regulator, which approves autonomous vehicle trials through exemptions under section 101 of the Road Traffic Act 1961. It actively encourages AV pilots in controlled settings such as business parks and ports, subject to safety and insurance conditions. That trial-friendly stance is one reason the city-state attracts deployments faster than tighter EU or US regimes.