Grab Holdings posted first-quarter revenue of $955 million Tuesday, up 24% year over year, as drivers using its Turbo AI assignment mode earned 23% more per online hour and quarterly profit climbed twelvefold to $120 million. The Singapore-based super-app published the results days after Jakarta capped ride-hailing commissions at 8%, down from 20%, the steepest regulatory shock in Grab’s 14-year history.
The juxtaposition mattered. CEO Anthony Tan spent much of Tuesday’s call arguing that Grab’s proprietary data foundation of more than 20 billion transactions is the moat that lets the company keep widening margins while a quarter of its mobility business gets re-priced by presidential decree. The pitch held on at least one desk: Morgan Stanley reiterated its $6.40 price target the same day.
What came through was a company sprinting to convert AI from a marketing slide into a per-trip line on the P&L, even as one regulator just rewrote the math on those trips.
The Headline Numbers
Q1 is Grab’s seasonally softest quarter, but every reported metric accelerated against last year. On-Demand GMV reached $6.1 billion, up 24%. Adjusted EBITDA hit $154 million, up 46%. Trailing twelve-month free cash flow finished the period at $489 million, per Grab’s first-quarter 2026 results announcement.
The deliveries unit pulled $510 million in revenue on $3.9 billion in GMV. Mobility booked $337 million, up 19%. Financial Services added $107 million, up 43%. Group monthly transacting users hit 52 million, with active driver-partners at an all-time high after 16% year-over-year growth.
CFO Peter Oey reaffirmed full-year guidance: 2026 revenue between $4.04 billion and $4.10 billion, with adjusted EBITDA of $700 million to $720 million. Grab also disclosed a planned acquisition of foodpanda’s Taiwan delivery business and confirmed approval of a cross-border Singapore-to-Malaysia ride-hailing license.
How Turbo Translates 14 Years Of Data
Tan opened the call with the line he wanted analysts to remember: drivers running Turbo earn 23% more per online hour than drivers who don’t. Turbo is an opt-in AI assignment system that picks routes, predicts wait times, and stacks orders against demand surges using Grab’s transaction history.
The benefit, in Grab’s framing, is that the AI is paid for by the work it absorbs. The driver moves more trips per hour without working longer. The platform earns its commission against a higher denominator. Both sides see a number that goes up.
- Turbo: route and order optimization for drivers, 23% hourly earnings lift for adopters in Q1 2026.
- Mai: a merchant chatbot from GrabX 2026, adopted by roughly half of single-store merchants, with adopters reporting a 15% sales increase.
- Carri: a delivery pickup robot, announced at GrabX, that handles the restaurant handoff so drivers don’t bleed minutes inside malls or lobbies.
The Indonesian Shock Grab Couldn’t Forecast
On May 1, President Prabowo Subianto signed Presidential Regulation No. 27 of 2026, capping ride-hailing platform commissions at 8% and lifting the driver share of fares to a minimum of 92%. The previous ceiling was 20%, per Indonesia’s official announcement of the new commission cap.
The decree caught the industry off guard. In its February Q4 call, Grab told investors no commission-cap changes were on the table. Indonesia is roughly 17% to 19% of Grab mobility GMV and around 20% of consolidated EBITDA, per Morgan Stanley’s note this week.
It’s essential, we believe, that together with regulators, we shape a balanced implementation of this decree, so that our Indonesian mobility marketplace remains healthy and that driver-partners’ earnings remain well supported.
That was COO Alex Hungate, speaking on Tuesday’s call. The quote signals the lobbying posture: engagement, not litigation. Grab is leaning on a possible carve-out from Prabowo’s speech, which named two-wheel ojol drivers specifically, leaving open whether the 8% applies to four-wheel ride-hailing as well.
That ambiguity is the difference between a small earnings nick and a structural margin reset. The Ministry of Transportation has not yet published implementation guidance, and Grab confirmed it is still in active talks with regulators.
The political backdrop is sharper than the headline number suggests. Tens of thousands of ojol drivers rallied at the presidential palace last summer, and according to the July 2025 ojol driver rally coverage, the protesters’ original demand was a 10% cap. Prabowo’s 8% figure went lower than the streets had asked for.
Mai And The Merchant Side Doing Quiet Heavy Lifting
Mai is Grab’s chatbot for restaurant and shop owners on the GrabMerchant app. Half of Grab’s single-store merchant base now uses it, and those adopters report 15% higher sales than non-users. The assistant pushes proactive nudges, drafts ad campaigns, edits menus, and answers operational questions on demand.
The deployment is the clearest case of Grab using third-party foundation models, then layering its own transaction data on top. Per Grab’s agentic AI rollout investor announcement, the merchant assistant runs on a stack that includes Anthropic’s Claude alongside other LLMs, with Grab’s own data layer on top.
Why this matters more than the headline lift: small merchants are the part of Grab’s marketplace most likely to churn during a downturn. A 15% sales bump for adopters is also a 15% retention argument for a population GoTo is fighting to win in the same markets.
Robotaxis On Punggol Streets, Robots In Mall Lobbies
On April 1, Grab and WeRide started Singapore’s first autonomous public ride service inside the Punggol residential estate. The Ai.R fleet, made up of WeRide GXR robotaxis and a Robobus, has logged roughly 30,000 kilometers of autonomous mileage with safety operators on board, per the Ai.R Punggol public launch press release. Rides are free until commercial pricing kicks in mid-2026.
Tan was careful on the call. Autonomy disrupting Grab’s human driver economy is a distant prospect, he said, and the more pressing automation play is closer to home: the wait time inside malls and office towers.
That is where Carri comes in. At Grab’s GrabX 2026 product event recap, Tan said drivers lose roughly 10% of their earning time inside buildings, finding the right counter or waiting for a customer to arrive. A robot that handles that handoff becomes free productivity that flows back to the driver.
If Indonesia’s 92% driver floor sticks, every minute of wait time clawed back is a minute the platform doesn’t have to subsidize through bonuses or surge pricing. Carri is small. The economic logic underneath it is not.
Loans Cross $1 Billion For The First Time
Grab’s financial services arm disbursed $1.1 billion in loans during Q1, up 67% year over year and the first quarter past the billion-dollar mark. The gross loan portfolio across GXS Bank in Singapore and GX Bank in Malaysia reached $1.4 billion, a 130% acceleration from the prior-year quarter.
Segment-level adjusted EBITDA loss narrowed to $17 million from $30 million. Combined deposits across the two digital banks held at $1.63 billion at quarter end, and revenue growth in the segment came primarily from lending rather than fee income.
Where Analysts Land
- $4.04B-$4.10B: Grab’s reaffirmed 2026 revenue guidance.
- $700M-$720M: 2026 adjusted EBITDA guide.
- $6.40: Morgan Stanley’s price target reiterated after the Indonesia decree.
- 12x: jump in Q1 net profit, $120M versus $10M a year earlier.
Morgan Stanley kept its Overweight rating and $6.40 target through the Indonesia announcement, which dropped GRAB shares roughly 6% on May 1. The reasoning, per the note: even if Indonesian mobility margin compresses, deliveries, financial services, and AI-driven productivity gains absorb the hit inside the existing 2026 guide.
The cleaner read on Tuesday’s results: Grab is leaning on AI-driven productivity to offset a regulatory shock that none of its forecasts had built in. The Turbo number is the proof of concept. The Indonesian decree is the stress test. Both arrived in the same quarter.
Bloomberg has reported that GoTo, Gojek’s parent and Grab’s chief Indonesian rival, said it would comply with the new rules while assessing operational and financial implications. The same calendar pressure now applies to both.
Frequently Asked Questions
When Does Indonesia’s 8% Commission Cap Take Effect?
President Prabowo Subianto signed Presidential Regulation No. 27 of 2026 on May 1, but did not specify an effective date. The Ministry of Transportation has yet to publish implementation guidance, and Grab confirmed on Tuesday’s call that it is in active talks with regulators on rollout timing. Until that guidance lands, the existing 20% ceiling continues to apply in practice.
Does The Indonesia Cap Apply To Four-Wheel Drivers Or Only Motorcycle Ojol Drivers?
The decree language references ride-hailing platforms broadly, but Prabowo’s announcement speech specifically named ojol two-wheel drivers, and protest organizers focused on motorcycle commissions. Grab and analysts including Morgan Stanley currently model the rule as primarily affecting two-wheel mobility, leaving the larger four-wheel ride-hailing segment less exposed pending Ministry of Transportation clarification.
How Does Grab’s Turbo Mode Actually Work?
Turbo is an opt-in driver mode that uses Grab’s transaction history to optimize trip assignment, sequencing, and routing in real time. The system stacks orders against demand surges, predicts pickup times, and reduces idle minutes between trips. Grab says drivers using Turbo earned 23% more per online hour than non-adopters during the first quarter of 2026.
When Will Punggol’s Ai.R Robotaxi Service Start Charging Riders?
Ai.R rides in the Punggol residential estate are free during the introductory phase that began April 1, 2026. Grab and partner WeRide plan to introduce paid commercial pricing in mid-2026, after usage and feedback data from the free window are processed. Safety operators remain onboard through the paid launch and into the discounted introductory fare period.
What Is Grab’s 2026 Revenue Guidance?
Grab reaffirmed full-year 2026 revenue guidance of $4.04 billion to $4.10 billion and adjusted EBITDA of $700 million to $720 million on Tuesday’s call. CFO Peter Oey said the targets remain firm despite Indonesia’s commission cap, with AI-driven productivity, deliveries growth, and lending expected to absorb the regional regulatory headwind.
How Is The Carri Robot Different From Existing Delivery Automation?
Carri is a Grab-branded pickup robot designed to handle the in-mall and in-lobby handoff between restaurant and rider, not the last-mile delivery itself. It targets the roughly 10% of driver earning time spent locating outlets and waiting for customers in tall buildings. Grab plans to test Carri in Southeast Asian malls before any wider rollout.
Tan’s framing on the call, that Grab is the system of record for Southeast Asian commerce, was always a bet that data scale eventually shows up in margins. Q1 made the number specific: 23% per online hour for the drivers running Turbo, 15% sales lift for merchants using Mai, twelvefold growth in net profit.
Whether that is enough to absorb a 92% driver share in Indonesia is the question every analyst will spend the next two quarters answering. The numbers reported Tuesday say Grab thinks it already has.
Leave a Comment