Community Discussion · Policy

Grab's AI Speeds Up 3x, Now What?

Mai Ken CaoMai Ken CaoAug 42026/08/04 358 views

Last night, while waiting for someone in the car, the center console screen was on, and I casually came across an interview with Grab CFO Peter Oey. He said AI had tripled Grab's product delivery speed, leading the company to raise its full-year guidance. I stared at the screen for a few seconds—not because of the 'triple' figure, but because of Grab itself. I used it when I went on a business trip to Singapore in 2019; back then, its delivery experience lagged behind domestic Meituan by at least two generations. Six years later, it turned things around thanks to AI.

2 replies

?
Ctrl + Enter to reply
Tiangong

Extreme weather is a major hurdle for AI implementation in Southeast Asia; Grab's models probably can't handle Bangkok's torrential rains either. However, its strength lies in heterogeneous data fusion. If it could integrate real-time weather and traffic conditions, it should be more stable than pure LSTM.

Long Yunfan

I just built a prediction model using LSTM last week, and it completely crashed when extreme weather hit... Can their predictions stay accurate even during Bangkok's heavy rainstorms?