Grab Pays $1.49 Billion for Atome to Leapfrog a Decade of Consumer Lending Buildout, Lifts 2028 EBITDA Target to $1.7 Billion
September 15, 2026 — M&A call on the acquisition of a 60% stake in Atome Financial
Grab Holdings used an investor call on September 15 to announce one of the largest deals in its history: a $1.49 billion cash purchase of a controlling 60% stake in Atome Financial, the Southeast Asian buy-now-pay-later and consumer lending platform. The transaction, structured with a performance-linked earn-out for the remaining 40% stake, is designed to compress what management repeatedly described as a decade of credit-model learning into a single deal, and it comes with a meaningful upgrade to Grab's group-level financial targets.
Group Targets Move Up Materially
The most consequential disclosure was not the deal itself but what it does to Grab's medium-term guidance. CFO Peter Oey raised the 2025-2028 revenue CAGR target to above 30% from 20%, and lifted the 2028 group adjusted EBITDA target to $1.7 billion from $1.5 billion. Of that $200 million increase, $60 million comes from the pending Foodpanda Taiwan deal, while $300 million is attributed to Superbank and Atome combined, an amount partially offset by a newly disclosed $160 million reinvestment into affordability and grocery/retail initiatives. Financial Services segment adjusted EBITDA, including Atome, is now guided to $500 million by 2028, up from a segment that was still loss-making by $110 million in 2025. The gross loan portfolio is expected to more than double, from just over $3 billion at the end of 2026 to over $6 billion by 2028.
Why Buy Instead of Build
Management was unusually direct about the economics behind the acquire-versus-build decision. Only 1% of Grab's 138 million annual transacting users currently borrow from the platform, and CEO Anthony Tan (referred to on the call as Ping Yeow Tan) said the combined platform could scale that lending user base by more than 10 times over the medium term. COO Alex Hungate framed the rationale bluntly: "This acquisition is an opportunity for us to leapfrog the 10 years of learning across multiple products and multiple markets that Atome has already been through. So we don't have to go through the pain of all the credit losses to train those models." Oey reinforced this with a chart showing that an organic build would sit meaningfully below Atome's existing scale for years, given that expected credit losses hit up front while loan revenue accrues over the life of the loan — a timing mismatch Atome has already worked through.
Atome's Scale and Credit Engine
Atome brings a gross loan book above $1 billion, more than 25 million cumulative transacting users, and relationships with over 30,000 merchant brands spanning travel, electronics, fashion, beauty and e-commerce — categories Grab's own F&B-heavy merchant base has never touched. The underwriting engine run by Atome uses over 100,000 variables and 131 models, with more than 22 live AI applications embedded across the risk stack; management said models are now built twice as fast, feature engineering is six times more efficient, and risk monitoring is four times more robust than before. GMV has grown ninefold over six years and fourfold over the past three years alone, with delinquency rates improving across every cohort — the 2026 cohort is said to be the best-performing yet. Average purchase frequency runs at an annualized rate in the 50s to 60s, reaching 62 transactions per user in the most recent quarter, a level of repeat engagement management pointed to as evidence that growth is coming from existing users rather than pure acquisition.
Deal Structure Limits Downside
Grab is paying the full $1.49 billion upfront for 60% control, funded entirely from existing cash, with the transaction expected to close by the third quarter of 2027. The remaining 40% stake is priced two years after completion within a valuation corridor of $2 billion on the floor and $4.5 billion on the cap, tied directly to Atome's adjusted EBITDA and revenue delivery over that window. Atome's existing management, led by founder and CEO Jefferson Chen, will continue running the business independently throughout the earn-out period. Hungate said the earn-out itself is the retention mechanism: "With this earn-out structure, I think there's a very strong incentive for them to remain around for those 2 years." On governance, Hungate said a Grab-controlled board subcommittee will set risk appetite parameters, with Atome's team executing within those guardrails, the same model already used with Grab's digital banks. Analysts pressed on potential conflicts given the earn-out structure, specifically whether Grab could use its cheaper cost of capital to refinance Atome's facilities and inflate the Phase 2 payout. Hungate said any such cooperation, including balance sheet or funding arrangements, would be conducted at arm's length as related-party transactions, "so there's no inflation of the earn-out because it will all be market-based arm's length."
The Ecosystem Data Argument
Grab leaned heavily on its proprietary data advantage as the strategic logic underpinning the deal. Hungate said the platform generates real-time signals from over 50 million ecosystem participants across more than 900 cities, with over 20 billion cumulative transactions since 2012 and 300 petabytes of data processed daily. The company disclosed 647 million cumulative repayment touchpoints, which it says allow its models to produce lower delinquency than traditional credit bureau models at comparable approval rates — relevant in a region where only 5% of adults own a credit card and just 14% have ever borrowed from a formal financial institution. Grab also cited concrete ecosystem uplift figures: drivers who take a loan complete 13% more rides per month on average, GXBank's FlexiCredit product in Malaysia drove an 11% GMV uplift on Grab after drawdown and a 54% uplift in spend through GXBank payment methods, and AI-driven lending recommendations to merchants lifted average loan disbursement by 19% in Thailand and 7% in Indonesia.
Profitability Timeline and Cost Levers
Grab reiterated that its Financial Services segment is on track to reach adjusted EBITDA profitability in the second half of 2026, up from a $110 million loss in 2025. Oey attributed the path to $500 million by 2028 to four levers: user growth from Atome's expanded footprint as the single largest driver, a shift toward longer-tenured, higher-yield products, declining credit provisions as underwriting models improve, and operating leverage as fixed underwriting and servicing costs are spread across a larger loan book. Atome's own loan portfolio skews short-duration, with an average tenor below six months across the entire book, according to Hungate.
Capital Returns Continue Alongside the Deal
Grab framed the acquisition as compatible with, rather than competing against, its buyback program. The company has repurchased $351 million through August 2026 under its 2026 authorizations and intends to complete approximately $900 million of the remaining mandate over the next 12 months. Cumulative repurchases since 2024 total $1.75 billion, representing over 10% of shares outstanding, all of which management intends to cancel. Oey said the company continues to target an 80% free cash flow conversion rate and will maintain an ample net cash buffer following the Atome transaction close.
Existing Products Stay in Place
On integration, Hungate confirmed that Grab's own pay-later products in Singapore and Malaysia, where they overlap with Atome's offering, will not be retired or merged, citing substantial white space in those credit markets: "It's really not a net sum zero situation between us and Atome, and we're both growing rapidly." Management also addressed bandwidth concerns given simultaneous integration efforts across Superbank, Stash and now Atome, with Tan noting each asset retains its own accountable leadership team and its own earn-out or standalone operating structure specifically to avoid straining a central integration team.