How Grab, Sea, and GoTo really make their money
Grab is still only at 9% financial-services revenue, but the pattern across Sea, GoTo, and Maya is clear: once the audience is built, the wallet line is the one that quietly takes over the P&L.
How much of Grab’s revenue comes from financial services?
Answer: About 9%: 253 million dollars out of 2.8 billion in 2024, up 44% year-on-year, with an EBITDA loss of 105 million dollars. The line is the fastest‑growing one at Grab, but it is still small and still loss‑making. One Southeast Asian unicorn already runs on financial services, two more are on the way there, and the rest are moving in the same direction whether they admit it or not.
At Sea Limited, the financial services arm earned 712 million dollars in adjusted EBITDA in 2024, about 4.6 times the e‑commerce arm that built it. The “payments and lending sidecar” is now more profitable than the marketplace itself.
The decade-long setup
The lines you see in 2024 earnings releases were planted 7 to 18 years ago. None of these companies woke up in 2024, decided to “do fintech,” and printed a profit line.
- Smart Money in the Philippines, the predecessor of what is now Maya. Wallet first. Bank licence in 2022.
- AirPay inside Sea / Garena to top up mobile games. The original use case was virtual goods, not commerce.
- GoPay inside Gojek, triggered by the problem of paying drivers in cash. Consumer wallet behaviour followed.
2016–2017. Lazada Wallet rolls out alongside Alibaba’s majority stake. A defensive build against Shopee and Grab. - GrabPay opens to third‑party merchants. Grab Financial Group formalises in 2018.
- Traveloka launches BNPL for flights and hotels. PayLater Card with BRI comes a year later.
- Sea rebrands its DFS suite to SeaMoney; AirPay migrates to ShopeePay. Bukalapak rolls out BukaDompet.
- Tokopedia merges into GoTo Financial.
- SeaMoney rebrands to Monee. Bukalapak shutters its physical‑goods marketplace and pivots to digital services.
Every one of these moves followed the same pattern. First, build or acquire an audience that does something often: ride, shop, book, play, top up. Then, layer a financial product into that existing behaviour. Drivers, gamers, travellers, marketplace buyers. The financial product was not a separate business. It was a new layer on top of an existing surface.
Where the money actually sits now
One of these companies already runs on financial services. Two are on the ramp. The rest are somewhere on the same curve. The slopes differ; the destination does not.
Sea Limited (FY 2024). Group revenue 16.8 billion dollars. Digital financial services revenue 2.4 billion dollars, up 35% year-on-year, on a 5.1 billion dollar loan book with a 1.2% 90‑day past‑due rate. DFS adjusted EBITDA is 712 million dollars, versus 156 million dollars for the e‑commerce business that sells the goods. The line that started as a top‑up tool for Garena games is now Sea’s most profitable segment after gaming itself.
Grab (FY 2024). Group revenue 2.8 billion dollars. Financial services 253 million dollars, 9% of revenue, growing 44%, with minus 105 million dollars in adjusted EBITDA. The unit is in build mode, not harvest mode. It is, however, growing more than twice as fast as mobility or deliveries. On current trajectories, the smallest, fastest‑growing line becomes either the profit engine or the largest write‑off on the P&L.
GoTo (FY 2024). Group gross revenue 18.1 trillion rupiah. Fintech segment 3.7 trillion rupiah, around 225 million dollars, up 95% year-on-year, with consumer loans outstanding up 172% in Q4. The fintech unit posted its first profitable quarter in Q4 2024. GoTo is running the steepest curve in the cohort.
Maya (FY 2024). 39 billion pesos in deposits, 68 billion pesos in loans disbursed, 5.4 million bank customers. Maya is the mirror image of Grab and Sea: a financial product that became a bank, rather than a marketplace that became a financial product. The strategic lesson is identical. Financial services compound when they sit on an existing trust relationship.
If you plot segment revenue and segment EBITDA as stacked bars for Sea, Grab, and GoTo, the pattern is clean. The financial services slice is a smaller share of revenue than of EBITDA. Revenue tracks adoption. EBITDA tracks underwriting discipline, pricing, and operational leverage. The financial layer monetises the surface more efficiently than the original marketplace does.
The case study everyone ignores
Bukalapak. The financial product did not save the marketplace. Physical‑goods sales fell below 3% of revenue. The company has been shutting down its core surface and pivoting to digital services and its Mitra warung network. The stock is down more than 85% since the 2021 IPO.
This is the unglamorous side of the same rule. A wallet compounds on top of an engaged audience. It does not create that audience from scratch. When the audience erodes, no wallet rescues it. When the daily habit leaves, the financial layer is a façade on an empty mall.
What Grab’s 9% should trigger in your head
Three practical points drop out of all this for operators who already have customers.
First, the math scales down. Grab’s effective take rate over on‑demand GMV is around 1.4%. Sea’s is around 2.4% over e‑commerce GMV. Both numbers live in the same 1–2% range that a mid‑market operator will see once they run the GMV × capture × take rate × float yield math on their own base. The fact that their revenue prints in billions does not mean the model needs unicorn‑only economics.
Second, the P&L timing is slow and lumpy. Sea’s financial services arm took roughly seven years to flip from cost line to profit engine. Grab’s line is still in the loss‑making build phase. GoTo just crossed into its first profitable quarter. The first three years of any financial‑product line are an investment period, not a return period. Teams that write the business case expecting Sea’s year‑seven margins in year one are budgeting fiction and will panic‑cut just before the curve steepens.
Third, the sequence is not decorative. Every successful player here launched a wallet before a card or a loan. The wallet creates three things at once: behaviour data, an always‑on balance to monetise, and a daily habit. Cards and loans are just ways of turning those three into revenue and EBITDA. Skip the wallet and the card underperforms. Push a loan into an audience you have not trained and it fails under load.
If you already have the audience, you are playing a different game
The 253 million dollar line inside Grab’s P&L is not a curiosity.
The 2.4 billion dollar line inside Sea’s P&L is not an exception.
The 3.7 trillion rupiah fintech line at GoTo is not a side project.
All of them started as tools stapled onto an existing surface: a driver wallet, a game top‑up, a BNPL for plane tickets. The difference between them and you is not the idea. It is the size and engagement of the audience they built before they shipped the first “finance” feature.
If you already have customers who come back often and trust you with operational workflows, the question is no longer whether you should build financial products. The questions are which product to launch first, how long you are willing to carry an investment P&L, and what kind of surface you are compounding on. For operators who already own an engaged surface, the real risk is not starting the clock.
If you already have the audience, every year you delay keeps you stuck at “Grab, before the 9%.”
