- Grab is rapidly expanding its physical AI capabilities, integrating robots and autonomous systems into its core operations across Southeast Asia.
- The company is testing drone deliveries in rural areas and autonomous delivery bots in urban centers to reduce delivery times and costs.
- Grab aims to reduce delivery times by up to 40% and cut operational costs through its AI-driven hardware expansion.
- The company is collaborating with automotive partners to develop self-driving ride-hailing vehicles in Singapore and other cities.
- Grab’s vision is to transform the superapp into a physical AI pioneer, building machines that deliver services to users.
In a sun-drenched corner of Grab’s Singapore headquarters, a delivery robot from a Japanese startup quietly navigates around employees sipping kopi, while a drone from a homegrown Malaysian firm hums through an indoor test corridor. This is no tech expo—it’s a typical Tuesday. Here, artificial intelligence isn’t just code on a screen; it’s whirring, rolling, and flying through the office corridors, learning the rhythms of human movement. The vision, led by Grab’s Chief Technology Officer, is audacious: to transform the region’s most dominant superapp into a physical AI pioneer. From automating last-mile deliveries to piloting self-driving ride-hailing fleets, the company is no longer just connecting people to services—it’s building the machines that deliver them.
Grab’s Multi-Pronged AI Expansion
Grab is rapidly deploying AI-driven hardware across Southeast Asia, integrating robotics and autonomous systems into its core operations. The company has launched pilot programs for drone deliveries in rural Philippines and Malaysia, where terrain and infrastructure limit traditional logistics. In urban centers like Jakarta and Bangkok, autonomous delivery bots are being tested for food and parcel drop-offs. According to Grab’s CTO, the goal is to reduce delivery times by up to 40% while cutting operational costs. The company is also collaborating with automotive partners to develop self-driving ride-hailing vehicles, with initial tests underway in Singapore’s designated autonomous zones. These efforts are part of a broader strategy to embed AI into the physical world—what the tech industry calls “physical AI”—transforming Grab from a digital platform into an embodied intelligence network.
The Road to Physical Intelligence
The shift toward physical AI didn’t happen overnight. Grab began as a ride-hailing app in 2012, focused on solving urban mobility gaps in underserved Southeast Asian markets. As it expanded into food delivery, financial services, and logistics, the limitations of human-dependent operations became clear—especially during the pandemic, when demand surged but labor shortages mounted. That crisis accelerated Grab’s investment in automation. By 2021, the company had established a dedicated robotics division within its engineering arm. It began acquiring AI startups specializing in computer vision and autonomous navigation. The turning point came in 2023, when Grab partnered with the Singapore government on a national smart mobility initiative, gaining access to real-world testing environments and regulatory sandboxes. This historical pivot—from digital intermediary to hardware-integrated platform—reflects a broader trend in tech, where giants like Amazon and Alibaba are also investing heavily in physical AI to control the full stack of service delivery.
The Minds Behind the Machines
At the helm is Grab’s CTO, a soft-spoken engineer with a background in robotics from MIT and stints at Google’s self-driving division. His philosophy is both pragmatic and competitive: “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.” This approach means developing core AI capabilities in-house (the “1”) while integrating and stress-testing third-party robotics (the “+n”). His team regularly benchmarks competitor bots on speed, obstacle avoidance, and battery efficiency. The motivation is not just innovation but survival—Grab operates in a fiercely competitive region, where rivals like Gojek and Foodpanda are pursuing similar automation strategies. By exposing their engineers to rival technologies, Grab cultivates a culture of adaptive learning, ensuring they stay ahead without becoming insular.
Impacts on Workers, Cities, and Rivals
The rise of physical AI at Grab carries profound implications. For gig workers, there is growing concern that automation could displace delivery riders and drivers. Grab insists that AI will handle high-volume, low-margin tasks, freeing humans for higher-value roles like customer support and complex deliveries. Urban planners, meanwhile, see potential benefits: reduced traffic congestion and lower emissions through optimized routing and electric-powered bots. However, regulatory challenges remain, particularly around safety standards and airspace management for drones. Competitors are reacting swiftly—Gojek has announced a $200 million AI mobility fund, while regional governments are drafting new frameworks for autonomous operations. For investors, the stakes are high: physical AI could significantly improve Grab’s profitability, long pressured by thin margins in its core businesses.
The Bigger Picture
Grab’s push into physical AI reflects a global inflection point where digital platforms are no longer content to mediate services—they want to execute them. This convergence of software, hardware, and urban infrastructure signals a future where AI doesn’t just recommend, but acts. In densely populated, rapidly urbanizing regions like Southeast Asia, the potential impact is enormous. If successful, Grab could serve as a blueprint for how tech companies integrate into the physical fabric of cities, reshaping how people move, eat, and live. But it also raises urgent questions about labor, equity, and control in an automated world.
What comes next is a new phase of experimentation and scaling. Grab plans to deploy hundreds of autonomous units across five countries by 2026, with full commercial operations targeted by 2028. The CTO remains cautious but optimistic: “We’re not replacing humans—we’re augmenting what’s possible.” Whether this vision delivers inclusive progress or deepens divides will depend not just on technology, but on how it’s governed. The robots are already in the office. Soon, they’ll be on every street corner.
Source: Fortune