Grab's H1 Profit Surges to US$355M as Its AI Layer Touches Every Part of the Superapp

When Grab reports numbers like these, it stops being just another Southeast Asian tech earnings call. The Singapore superapp posted US$1.95 billion in revenue for the first half of 2026, up 23 percent year on year, with profit of US$355 million — a leap from US$30 million in the same period last year. The jump matters less for the headline than for what drove it: an AI layer that Grab says is now embedded across every part of its platform, from matching drivers to pricing deliveries to underwriting loans.

The second quarter alone tells the story of a company compounding on itself. Revenue hit US$997 million, up 22 percent, while adjusted EBITDA rose 54 percent to US$168 million — the eighteenth consecutive quarter of adjusted EBITDA growth. On-demand gross merchandise value climbed 21 percent to US$6.5 billion, and monthly transacting users hit a record 54 million, up 17 percent. Financial services was the standout: revenue grew 59 percent to US$134 million, with loans disbursed rising 72 percent to an all-time high of US$1.2 billion, helped by the consolidation of Indonesian bank Superbank in June, which also contributed a one-time US$307 million gain to Q2 profit.

The AI thread running through these numbers is deliberate. Grab has spent years building its own machine-learning stack — demand forecasting for deliveries, dynamic pricing for rides, fraud detection for its fintech arm — and the company now frames that work as a single "AI intelligence layer" that strengthens as it scales. Management raised full-year guidance to US$4.10–4.15 billion, growth of 22–23 percent, a signal that the AI-driven efficiency gains are expected to keep compounding through the second half.

For Singapore, Grab's results are a proxy for the health of the region's digital economy — and evidence that the country's flagship consumer internet company is betting its next chapter on AI rather than just ride-hailing share. The record 54 million transacting users and the 14 percent rise in average merchant earnings suggest AI is improving the marketplace for participants, not just extracting more from them, which is the test most platforms still fail.

Why it matters for Singapore: Grab is the closest thing the country has to a national AI flagship — a Singapore-founded company using machine learning at consumer scale across six countries. Its ability to convert AI investment into actual profit growth, quarter after quarter, is one of the strongest local data points that the AI build-out translates into earnings, and it sets the tone for how investors value the rest of the SG tech ecosystem.