Why self-driving cars won't take off in Indonesia anytime soon, according to Grab

While self-driving cars might feel like the future elsewhere, Grab doesn’t expect them to dominate Indonesian roads anytime soon. Speaking at the Asia Economic Summit 2025, Grab COO Alex Hungate (pictured) laid out a grounded vision for autonomous vehicles (AVs) in Southeast Asia - tempering the hype with a dose of local realism.

“Maybe Singapore might be an exception,” Hungate said. “But the major cities in Southeast Asia will be slower [to adopt AVs] for three reasons.”

At the top of his list: complexity. Urban environments in countries such as Indonesia present vastly different challenges than the controlled settings where AVs are being tested in the US or China. Narrow roads, chaotic traffic flows, and the prevalence of two-wheeled vehicles - especially motorbikes - make safe navigation a technical and regulatory minefield.

Don't miss: Grab launches AI Centre of Excellence to drive SEA innovation

“The majority of rides in Indonesia on Grab are on two wheels, not four wheels,” Hungate noted. “So that’s one of the complications…” Some companies are already experimenting with self-driving two-wheelers, but this presents its own set of challenges, Hungate added.

In developed markets, motorbikes are typically subject to strict lane discipline and are often prohibited from weaving through traffic. As a result, AVs designed for such environments may struggle when applied to countries such as Indonesia, where motorbikes flow fluidly between vehicles, often unpredictably and with minimal regard for lane markings. This style of movement is notoriously difficult for autonomous systems to replicate or safely respond to. 

“I think that would be a really big challenge for AVs for a long time to come,” he emphasised.

Beyond infrastructure, the economics of AV deployment in Southeast Asia don’t yet stack up. Hungate pointed out that the cost curve - which favours AVs in higher-income markets - is still out of reach for most of the region. “The business rationale for bringing them in will be less strong,” he said.

A measured approach to the inevitable

Grab isn’t dismissing AVs outright. The company sees their eventual rise as a matter of “when”, not “if” - but insists that any transition must benefit its community of driver-partners.

“We want to study how this will change employment opportunities for drivers,” said Hungate. “New types of jobs will be created… remote drivers, servicing the vehicles. We want to understand what those are.”

Grab plans to be among the first to experiment with and optimise hybrid models that combine AVs and traditional ride-hailing. Still, the message was clear: Indonesia and most of Southeast Asia are in for a long, gradual adoption. “There is going to be a long tail of transition,” Hungate said.

AI, the clearer bet

While AVs sit at the far end of the spectrum, Grab is focused on the near-term frontier - AI that drives real impact today. Much of the company’s recent US$1.5 billion convertible bond will be channelled into AI efforts. “We will use it to double down on AI,” Hungate said, pointing to partnerships with OpenAI, Anthropic, and others.

Key to Grab’s strategy is local adaptation. Early experiments with voice recognition models showed dismal results in understanding Southeast Asian dialects - just 45% accuracy. “We’ve now fine-tuned it with local dialects and we’re getting up to 87%,” Hungate said, adding that Indonesia will benefit from continued investment in covering local sub-language groups.

Access is one of the first goals. Grab is working with associations for the visually impaired to test voice-based interfaces. But Hungate sees a broader behavioural shift on the horizon: “Perhaps the way in which we all interact with apps will dramatically change.”

AI is also reshaping Grab’s operational core. Drivers are now supported by a real-time co-pilot system that recommends optimal positioning for higher earnings. On the food delivery side, “hyper-batching” ensures multiple orders can be handled in a single trip without delays - boosting productivity and pay.

For merchants, Grab has developed an AI-powered chatbot that acts like an all-in-one advisor - part CFO, part marketer, part emotional support. “We have now 80,000 merchants using this,” said Hungate. “You see the chat going back and forth - ‘Hey sis, it’s been a tough weekend.’ ‘Don’t worry, sis. Keep going. This week’s much better than last week.’” Grab didn’t set out to build a chatbot that feels like a friend - but in practice, Indonesian merchants are treating it like one.

The US$100 million sprint

To future-proof its workforce and rewire its internal thinking, Grab went as far as pausing business-as-usual operations for nine weeks last year in what it dubbed a “generative AI sprint.” Engineers stopped shipping features and instead focused solely on experimenting with AI.

“It’s kind of a crazy thing to do for a company of our scale,” Hungate said. “The cost of that is probably just our engineering time for nine weeks - about US$100 million.”

The result: an internal rearchitecture of Grab’s entire tech stack, more AI-driven project pipelines, and a pivot in hiring towards data science and advanced AI talent.

Despite advances in automation, Hungate emphasised that practical, human-centred applications remain critical. “AI has to solve real-world problems… in a human-centric way,” he said. While AVs may take time to scale in Southeast Asia, Grab is not waiting around - focusing instead on AI and platforms that meet the region’s immediate needs.

Related articles:
Grab, OVO roll out pilot nutrition programme for special needs schools
Grab, OVO expand digital push to Pontianak through #KotaMasaDepan programme
Grab takes to the skies: Philippines launches drone delivery pilot