Grab's Core Is Compounding While the Stock Sits Near Its Low

Later this quarter, Grab will let passengers in the Singapore neighborhood of Punggol do something new: open the Grab app and book a driverless vehicle on demand, riding directly between points along its two shuttle routes instead of around a fixed loop. The on-demand service rolls out to the public in the fourth quarter, with commercial fares to follow, and it builds on the Ai.R autonomous ride service Grab launched with WeRide in the area back in April — Singapore's first-ever autonomous public ride service. It is real, and it is almost certainly the least important thing about this stock right now.

The stock trades near $3.02, down roughly 40% for the year and a hair above its 52-week low of $2.96 against a high of $6.62. A growth headline like this, landing on a beaten-down name, reads like a catalyst. It isn't. Scaling driverless coverage across one housing estate — passenger counts in the thousands, mileage in the tens of thousands of kilometers — has no bearing on a company that books nearly $4 billion of revenue a year. The AV program is brand muscle and a call option on a driverless future, not current economics. Anyone paying up for the Punggol pilot is betting on the wrong line item.

The core is compounding while the stock isn't

The actual disconnect is elsewhere, and it is the kind the market gets wrong. In the very quarter before this headline, Grab's core numbers were compounding. Second-quarter revenue rose 22% year over year to $997 million. Adjusted EBITDA jumped 54% to $168 million, a 16.9% margin and the eighteenth straight quarter of growth. Monthly transacting users hit a record 54 million, with on-demand gross merchandise value up 22% on a constant-currency basis. Management used that strength to raise full-year 2026 guidance to $4.10–4.15 billion of revenue and $720–740 million of adjusted EBITDA, and — already sitting on about $5.4 billion of net cash — authorized another $750 million of share repurchases.

Now put a multiple on it. At a market cap near $12 billion and that net cash on the balance sheet, the raised guidance works out to roughly 10x forward EV/EBITDA. Pair that with an EBITDA line growing at a 40%-ish pace, and the stock is being priced for a slowdown the company's own guidance does not show. That is the definition of a valuation-versus-growth disconnect: the price is not the problem, the narrative is.

The bear case has a number behind it

But acting like this is pure panic would be lazy, and it would be the kind of reflex that gets people hurt. Adjusted EBITDA is a management number. It strips out items like stock-based compensation. On a plain accounting basis, Grab generated negative operating cash flow and negative free cash flow over the trailing twelve months. That gap — a company reporting record EBITDA growth while cash from operations has not yet turned positive — is the factual kernel inside the market's skepticism. This is still a grow-the-top-line-first, convert-to-cash-later story, and until the cash conversion lands, the discount has a basis.

That is how a reasonable investor should file the Punggol news. Read it as proof Grab can deploy autonomy in a working regulatory environment, not as a new revenue line and not as the reason to own the stock. The number that actually decides whether this cheap-looking grower re-rates is whether the EBITDA finally shows up as operating cash. The day it does, a business growing revenue over 20% and EBITDA near 40% at roughly 10x forward EBITDA is cheap by almost any math — but the timing of that conversion, not a driverless shuttle in one neighborhood, is what makes or breaks the case.