Grab's Record Quarter Leans on a One-Off Superbank Gain
Revenue and EBITDA growth are real, the US$235m profit headline needs an asterisk, and the lending book is now the number to watch.
Grab reported its second-quarter 2026 results on 4 August. Revenue rose 22% year on year to US$997m, adjusted EBITDA grew 54% to US$168m, and profit for the period came in at US$235m, against US$20m a year earlier. The company raised its full-year guidance to US$4.10bn to US$4.15bn in revenue and US$720m to US$740m in adjusted EBITDA, and authorised another US$750m of buybacks, taking the cumulative total since 2024 to US$1.75bn.
Those are good numbers. One of them deserves a closer look.
Read the footnote before the headline
The earnings release filed with the SEC says the quarter's profit includes a US$307m gain recognised when Grab began consolidating Superbank, one of its digital banks, in June. Grab itself calls the remeasurement gain one-time in nature. On simple arithmetic, a US$235m profit that contains a US$307m accounting gain means the underlying bottom line was not positive this quarter. Management also warned that second-half profit would keep moving around because of fair value measurements.
That does not make the quarter weak. Adjusted EBITDA, which excludes the gain, has now grown for 18 consecutive quarters, and margin widened to 16.9% of revenue from 13.3% a year ago. Adjusted free cash flow over the trailing twelve months was US$450m. But anyone using "Grab made US$235m" as evidence that Southeast Asian super-apps have finally become solidly profitable on a GAAP basis is reading the headline, not the filing.
The engine underneath
The operating detail is more useful than the profit line:
- On-demand GMV reached US$6.5bn, up 21%. Deliveries GMV was US$4.25bn and mobility US$2.21bn, with mobility transactions up 28%. Rides growing faster than ride value suggests cheaper trip types are bringing in volume.
- Monthly transacting users reached a record 54 million.
- Incentives were US$706m in the quarter, with on-demand incentives at 10.9% of GMV. That is still a lot of money spent to keep both sides of the marketplace moving.
The part I would watch most closely is financial services. Revenue there rose 59% to US$134m. The gross loan portfolio was US$2.3bn, up 197% year on year. Disbursements hit a record US$1.2bn in the quarter, and customer deposits across Grab's three digital banks, including GXBank here in Malaysia, came to US$2.5bn. The segment still lost US$15m at the adjusted EBITDA level, an improvement from US$26m.
Tripling a loan book in a year is the kind of growth that looks great until a credit cycle turns. Grab's edge is that it can see drivers' and merchants' earnings in real time, which is better underwriting data than most banks have. The risk is concentration: the same people who earn through Grab borrow from Grab, so a slowdown in rides or orders hits income and repayment together.
What it means for builders here
For Malaysian founders, Grab is a partner, a distribution channel and a competitor, often all at once. Three things stand out.
First, CEO Anthony Tan said in the release that the Grab "intelligence layer" now runs through every part of the platform. Merchant tooling, pricing and dispatch are increasingly model-driven. If you sell software to F&B or retail merchants, assume Grab will bundle more of it for free.
Second, the balance sheet is enormous. Gross cash liquidity was US$7.4bn, and Grab is spending on buybacks and acquisitions, including Stash Financial in the US, completed in July. It can outlast any local competitor in a price war.
Third, GXBank's deposit and lending growth is the closest thing Malaysia has to a live test of whether a platform bank can out-underwrite incumbents. Fintech builders should watch its credit quality more closely than Grab's share price.
What to watch next quarter: how Superbank's numbers look once consolidated without the one-off gain, whether loan growth slows or credit costs rise, and whether incentives as a share of GMV keep edging down.
Sources