Grab’s $1.49B Atome Deal Is a Bet on Debt, Not Deliveries
Ride-hailing and food delivery are nice businesses. Lending people money is where Grab thinks the serious profit is — and it has just put $1.49 billion on the table to prove it.
Grab didn’t spend $1.49 billion to become better at delivering noodles. It spent it because the real money in a super-app is lending to the bloke ordering the noodles — and the merchant selling them.
On September 15, Grab agreed to buy a controlling 60% stake in Singapore-based Atome Financial for $1.49 billion in cash. This is not a cute fintech bolt-on. It is Grab putting a large, very deliberate wager on consumer credit becoming the profit engine beneath Southeast Asia’s apps.
And before anyone starts clapping along to the usual “financial inclusion” soundtrack, let’s call it what it is: a high-upside bet with proper credit risk attached. Grab is buying speed, underwriting data and distribution. It is also buying a front-row seat to every bad decision consumers make when easy instalments meet a tough economy.
Grab is paying for a shortcut that would take years to build
Atome Financial operates across Singapore, Malaysia, the Philippines, Indonesia and Thailand. Its products include buy now, pay later loans, consumer cash loans, BNPL cards and digital lending. Grab says the business has served 25 million cumulative transacted users and brings a merchant network of more than 30,000 brands.
That is the asset here. Not the shiny app. Not the BNPL branding. The asset is a working credit machine in five markets, with borrower behaviour, repayment patterns, fraud signals, collection processes and regulatory know-how already baked in.
Building that internally would be expensive and slow. Worse, it would involve making a mountain of small lending mistakes before your risk models get remotely useful. Grab already has payments, partner lending, insurance and digital banks in Singapore, Malaysia and Indonesia. But its consumer-credit footprint has not matched its mobility and delivery reach.
Atome fills that hole quickly.
Grab has nearly 54 million monthly transacting users in its ecosystem. Atome brings consumers who borrow and a large merchant base that wants customers to spend more. That creates the classic platform flywheel: rides and food orders create transaction data; transaction data supports credit decisions; credit encourages spending; spending makes the platform more valuable to merchants; more merchants create more transactions.
It sounds wonderfully neat in a presentation. In reality, it only works if the credit losses stay controlled. That is the whole game.
Grab says Atome’s roughly $1 billion gross loan portfolio has maintained disciplined credit quality, with delinquency rates improving or stable across borrower cohorts. Fine. That is encouraging, but operators and investors should remember what it is: a snapshot, not a guarantee. Credit books often look smartest just before the economy gives them a proper test.
The clever bit is not the $1.49 billion. It is the remaining 40%
The first phase gives Grab control: 60% of Atome Financial for $1.49 billion in cash, including $260 million of primary growth capital. The deal is expected to close by the third quarter of 2027, subject to regulatory approvals and customary conditions.
Then comes the part more founders should study.
Grab has agreed to buy the remaining 40% about two years after the initial close, but not at a fixed price. The eventual valuation will be tied to Atome’s demonstrated performance. The formula gives a 75% weighting to annualised adjusted EBITDA at a 13.0-times multiple and a 25% weighting to annualised revenue at a 2.5-times multiple.
The final equity valuation is capped at $4.5 billion and floored at $2 billion. At least half of the second-stage consideration will be paid in cash.
That is a grown-up deal structure.
Grab gets control now and consolidation of the business after closing. Atome’s management keeps running the operation. The sellers get meaningful upside if the business genuinely performs. But Grab does not blindly hand over the full victory cheque before the race has been run.
Too many acquisitions go wrong because the buyer pays tomorrow’s results today, in cash, while the operators who built the target take their chips off the table. Grab has avoided the dumb version of that problem. It has made part of the price contingent on earnings and revenue that Atome must actually produce.
There is still risk, of course. Any formula can create incentives to optimise the metric rather than the business. A management team facing a two-year earn-out may push growth too hard, underinvest in long-term controls, or favour accounting-friendly decisions. Grab will need brutal clarity around credit standards, provisions, compliance and customer outcomes — not just a nice EBITDA number at the finish line.
But the structure is miles better than paying a fixed premium for a forecast written by investment bankers with a straight face.
Why lending matters more than rides and food
Grab is a super-app, which is another way of saying it is trying to own more of the consumer’s daily spending. Mobility and deliveries create frequency. They get people opening the app. But these businesses are operationally demanding, fiercely competitive and burdened by drivers, discounts, logistics and local market economics.
Financial services can be different. Done well, lending has higher margins and fewer physical moving parts. Done badly, it becomes a very efficient way to torch capital.
Reuters reported that Grab sees Atome helping it leapfrog the expansion timeline for BNPL in the Philippines, Indonesia and Thailand. That is exactly why this deal matters. Grab is no longer treating finance as a supporting feature for rides and deliveries. It is treating it as a standalone growth pillar.
The company has upgraded its 2028 targets to $1.7 billion in adjusted EBITDA and group revenue growth above 30% annually from 2025 to 2028. It expects its Financial Services segment, including Atome, to generate $500 million in adjusted EBITDA and build a combined gross loan portfolio above $6 billion by 2028.
That would be a substantial jump from Grab’s consumer lending book of $2.3 billion at the end of June 2026.
I would not take those targets as gospel. Targets are management’s job; delivery is their test. But the direction is unmistakable: Grab wants finance to carry more of the company’s future economics.
The overlooked angle: this is a funding-cost play
Most commentary will focus on BNPL, AI underwriting or financial inclusion. All relevant. The more interesting point is funding.
Consumer lending is not magic. You lend money at one cost and need to get it back at a higher return, after defaults, fraud, collections, regulation and overhead. The cost of money matters enormously.
Grab operates digital banks in Singapore, Malaysia and Indonesia. Reuters reported that Grab believes those banking operations can lower the cost of funding for Atome’s assets. If that works, Grab is not merely adding borrowers. It is tightening the economics underneath every loan.
That is where platform strategy becomes real strategy. The app provides distribution. The lending engine provides underwriting. The digital banks may provide cheaper funding. The merchant network provides places to spend. Each piece makes the others better.
Or at least that is the theory. The risk is that “ecosystem” becomes a fancy word for concentration. If consumers, merchants, loans and funding become too tightly linked inside one platform, a downturn can hurt several parts of the machine at once.
The uncomfortable truth about BNPL
BNPL is often marketed as a friendlier alternative to old-school credit. Sometimes it is. Spreading the cost of a purchase can be sensible. For people outside traditional banking, access to transparent credit can be materially useful.
But let’s not pretend every instalment plan is an act of charity.
The same frictionless experience that helps a responsible borrower manage cash flow can help an overstretched borrower spend money they do not have. And the data that improves underwriting also makes it easier to target people with increasingly precise offers.
Grab says 68% of its driver-partner borrowers accessed formal credit for the first time through Grab in 2025, and half said they did so to avoid predatory lenders. That is a meaningful opportunity. It also creates a serious obligation. When you are lending to people with limited formal credit histories, your collections, disclosures and product design matter just as much as your clever algorithms.
The winner here will not be the company that approves the most loans. It will be the company that earns the right to make the second, third and tenth loan without wrecking the customer or its own balance sheet.
What this means for you
If you are a founder, steal the deal structure, not the headline. Grab is paying real money, but it is tying future value to future performance. When buying a business, do not pay a fantasy price for forecast revenue. Use staged ownership, earn-outs and clear operating control where appropriate. Protect the downside without insulting the people you need to keep building.
If you are an operator, understand that distribution is increasingly more valuable than a feature. Atome did not just build lending tools. It built a credit business that plugs into Grab’s huge consumer and merchant network. Ask yourself: does your product become dramatically better when connected to someone else’s customer base, data or funding? If yes, you may be a strategic asset — not merely a vendor.
If you are an investor, do not get hypnotised by “AI-powered lending” or a $6 billion loan-book target. Watch credit quality, funding costs, provisioning, regulatory approvals and whether the promised cross-sell actually turns into profitable behaviour. Loan growth without disciplined losses is not growth. It is delayed bad news.
And for everyone else: remember that the best acquisitions do not buy revenue. They buy time, capability and a position that would be brutally difficult to build from scratch. Grab has bought all three. Now it has to prove it can lend billions without losing its head.