What happens when a superapp built on millions of human drivers starts betting on robots? For Grab, Southeast Asia’s ride-hailing and delivery giant, the answer is already rolling through the corridors of its Singapore headquarters—and soon, through the streets of a public housing estate.
On May 20, Grab announced that one of its robots, named Carri, will begin making deliveries in Singapore’s Punggol district, a designated hub for testing robotic services. But Carri has already been quietly navigating the hallways of Grab’s own office for months. And here’s the twist: it’s not the only robot in the building.
Why This Matters Right Now
This isn’t just a tech demo. Grab’s push into physical AI—robots that move, sense, and interact with the real world—signals a fundamental shift in how the superapp sees its future. With driver costs rising, urban congestion worsening, and competition intensifying from rivals like Gojek and Shopee, automation is no longer a distant possibility. It’s becoming a strategic necessity.
For millions of users across Southeast Asia who rely on Grab for food, groceries, and rides, this could mean faster deliveries, lower costs, and new services. But it also raises questions about what happens to the human drivers who are the backbone of the platform today.
How Grab’s Physical AI Strategy Unfolded
Suthen Paradatheth, Grab’s chief technology officer, has been with the company almost since its founding. He’s seen the superapp evolve from a ride-hailing startup into a sprawling ecosystem spanning transport, food delivery, payments, and logistics. Now, he’s overseeing what he calls the “physical AI” chapter.
“We don’t oblige our business units to just use our robots,” Paradatheth told Fortune during an interview at the Asia Tech (ATx) summit. “If you go to the Grab office now, you’ll see robots from other companies as well. We use a 1+n strategy which keeps us on our toes.”
That “1+n” approach means Grab builds its own robots—like Carri—but also tests and deploys machines from competitors. It’s a deliberate strategy to avoid tunnel vision and ensure the company is always benchmarking against the best in the world.
Who Is Affected and What Officials Are Saying
The immediate impact is on Grab’s delivery operations. Carri is designed to handle last-mile deliveries in dense urban environments, freeing human drivers to focus on more complex or high-value trips. In Punggol, the robot will navigate pedestrian walkways and public spaces, delivering food and parcels to residents.
Paradatheth emphasized that the goal is not to replace drivers but to augment them. “Robots can handle the predictable, repetitive routes,” he explained. “That allows our driver-partners to earn more by focusing on trips that require human judgment and interaction.”
Still, the long-term implications are hard to ignore. If physical AI scales successfully, the balance between human and machine labor on Grab’s platform could shift dramatically.
What We Know So Far — and What Remains Unclear
What we know:
- Carri will begin commercial deliveries in Punggol district starting May 20.
- Grab uses a “1+n” strategy, testing robots from multiple vendors including competitors.
- The company sees physical AI as a core part of its future, alongside automated driving.
What remains unclear:
- How quickly the robot fleet will expand beyond Punggol.
- The exact cost savings and efficiency gains Grab expects.
- How driver-partners will be affected in the medium to long term.
- Regulatory hurdles for wider deployment across Southeast Asia.
Risks, Concerns, and the Balanced View
Grab’s physical AI push is not without risks. The technology is still nascent, and public acceptance of autonomous robots in public spaces varies widely. Technical failures, safety incidents, or privacy concerns could derail adoption.
There’s also the human cost. While Grab frames automation as a tool to boost driver earnings, critics argue that it could eventually displace workers. The company will need to navigate this tension carefully, especially in markets where ride-hailing drivers rely on the platform for their primary income.
On the flip side, the “1+n” strategy is a smart hedge. By testing competitors’ robots, Grab ensures it isn’t locked into a single technology stack. If one approach fails, another can take its place.
Why Similar Trends Are Growing Across Southeast Asia
Grab is not alone in betting on physical AI. Across the region, logistics companies, e-commerce platforms, and food delivery services are experimenting with autonomous robots and drones. Singapore, with its supportive regulatory environment and advanced infrastructure, has become a testing ground.
In China, Meituan and Alibaba have already deployed delivery robots in select cities. In Japan, ZMP and other firms are testing autonomous carts. Grab’s move positions it to compete not just regionally but globally in the race to automate last-mile logistics.
“We use a 1+n strategy which keeps us on our toes.” — Suthen Paradatheth, CTO, Grab
What Readers, Users, or Investors Should Know Now
For everyday Grab users, the immediate impact will be subtle. You might see a Carri robot delivering your food in Punggol, but the service will feel largely the same. Over time, however, expect faster delivery times and potentially lower fees as automation reduces operational costs.
For investors, Grab’s physical AI strategy is a signal of long-term thinking. The company is investing in technologies that could give it a durable competitive advantage—if executed well. The “1+n” approach reduces technology risk, but execution risk remains high.
For driver-partners, the message is mixed. In the short term, automation could mean more efficient routes and higher earnings. In the long term, the role of human drivers may evolve toward supervision, maintenance, and customer service rather than driving.
What Could Happen Next
If the Punggol trial succeeds, expect Grab to expand Carri’s deployment to other parts of Singapore and eventually to other Southeast Asian cities. The company is also likely to deepen its investment in automated driving technology, which Paradatheth described as a complementary piece of the physical AI puzzle.
Partnerships with robot manufacturers, both in-house and external, will become more critical. Grab may also explore white-labeling its robot technology to other logistics providers, creating a new revenue stream.
Our Take: Why This Story Matters Beyond One Incident
Grab’s physical AI strategy is a microcosm of a larger shift happening across the tech industry. The line between digital platforms and physical infrastructure is blurring. Companies that once operated purely in software are now building hardware, deploying robots, and managing real-world logistics.
What makes Grab’s approach interesting is the humility embedded in the “1+n” strategy. By using competitors’ robots, the company acknowledges that no single player has all the answers. It’s a lesson in strategic flexibility that other superapps would do well to learn.
For now, the robots are rolling. And if Grab’s CTO has his way, they’ll keep rolling—alongside human drivers, not instead of them.
FAQs
What is Grab’s physical AI strategy?
Grab is investing in robots and automated driving technology to handle last-mile deliveries and logistics. The company uses a “1+n” approach, building its own robots while also testing machines from competitors to stay competitive.
Why does Grab’s CTO use competitors’ robots in the office?
Suthen Paradatheth says the “1+n” strategy keeps the team on their toes. By testing robots from other companies, Grab ensures it is always benchmarking against the best and avoids becoming complacent with its own technology.
Will Grab’s robots replace human drivers?
Grab says the goal is to augment, not replace, human drivers. Robots handle predictable, repetitive routes, allowing driver-partners to focus on more complex trips that require human judgment. However, long-term implications remain uncertain.
Where is Grab deploying its delivery robots first?
Grab’s robot, named Carri, started deliveries in Singapore’s Punggol district on May 20. Punggol is a designated hub for testing robotic services in the city-state.
