Anthony Tan agreed on 15 September 2026 to spend US$1.49 billion of Grab’s cash on a lending business he does not fully own yet. The deal buys Grab’s founder and chief executive 60 percent of Atome, a Singapore instalment-loan platform (buy now, pay later, or BNPL: the customer takes the phone or the flight today and pays it off in instalments, interest-free if every payment lands on time) operating across five Southeast Asian countries. The other 40 percent arrives roughly two years later, at a price nobody can fix today.
In 2011, a classmate at Harvard Business School needled Tan about his own family. His great-grandfather had driven a taxi in Malaya. His grandfather turned that into Tan Chong Motor, one of Malaysia’s biggest car distributors, and his female friends back home still could not flag a taxi in Kuala Lumpur after dark without worrying what happened next.
Tan and a course-mate, Hooi Ling Tan, entered a venture-plan competition with an idea for an app that would send a driver straight to a rider’s exact spot. They placed second and won US$25,000, added their own savings and a loan from Tan’s mother, and launched MyTeksi in Kuala Lumpur in June 2012.
The company that grew out of that prize money is now Southeast Asia’s biggest ride-hailing and delivery platform, worth about US$12 billion on Nasdaq.
Tan does not own much of it. When Grab listed through a SPAC merger in December 2021 (a SPAC is a shell company already trading on the stock market that buys a private company to take it public, skipping the usual IPO process), its board gave his shares, and shares his two co-founders agreed to vote through him, 45 times the voting power of an ordinary share, the kind of structure founders now write into the deal before they ever ring a bell.
The arrangement handed him 60.4 percent of the vote against a 2.2 percent economic stake at listing.
That was already more voting power than SoftBank, Grab’s largest outside shareholder, which held 18.6 percent of the stock and just 7.6 percent of the vote.1
By January 2026 Grab put his share of the vote at 59.1 percent, proxies included, and asked shareholders to double each Class B share from 45 votes to 90, which would take him to as much as 74.9 percent.
That July he sold 93 percent of what he still held directly: 400,000 shares for about US$1.6 million, on a trading plan filed with regulators months in advance so a chief executive’s own trades cannot be timed off inside information. It left him 28,498 Grab shares in his own name.
He now runs a US$12 billion company on votes, not equity, and he just committed US$1.49 billion of it to a fintech he cannot yet fully control.
The money is affordable because Grab spent a decade losing it and then stopped. It lost US$3.6 billion in 2021, much of that a one-time accounting charge tied to the listing, and another US$1.74 billion in 2022. Four years of quarter-by-quarter cuts later, that became a US$235 million profit in the quarter that just closed, Grab’s eighteenth straight quarter of rising adjusted profit.
You’ve just seen the scale and discipline Atome already has, and this is exactly why we chose to acquire rather than build this ourselves.
Atome brings what that decade of losses did not buy: a loan book of about US$1 billion, 25 million paying customers across Singapore, Malaysia, Indonesia, the Philippines and Thailand, ties to more than 30,000 merchant brands, and an underwriting model built to price credit for the roughly seventy percent of Southeast Asian adults banks will not lend to. “You’ve just seen the scale and discipline Atome already has, and this is exactly why we chose to acquire rather than build this ourselves,” Tan told investors on the call announcing the deal.
Building that book from zero, one repayment history at a time, would have cost Grab years it did not have.
So Tan structured the bill to match the risk.
Grab pays US$1.49 billion now for 60 percent, of which US$260 million is fresh growth capital going into Atome itself. The remaining 40 percent is not due for two more years, and its price is a formula, not a number: thirteen times Atome’s annualised earnings weighted 75 percent, plus 2.5 times its annualised revenue weighted 25 percent, floored at US$2 billion and capped at US$4.5 billion.
If Atome performs, the second cheque is larger and Grab already knows the ceiling. If it does not, Grab has not paid up front for growth that never showed up.
The version of this move available to a founder without Grab’s balance sheet is the same shape at a smaller scale. When a rival or an adjacent operator already holds a licence, a merchant network, or years of repayment data that would take a long time to build from nothing, buying the controlling stake and deferring the rest against a formula tied to real performance costs less, in cash and in time, than starting from zero or paying full price today for growth that has not happened yet.
Tan is due to keep selling down what little Grab stock he holds directly, on the same pre-set schedule that took him from 400,000 shares to 28,498 in one afternoon in July, before September 2027.
Votes, not shares, are what keep him in the chair.
Votes, not shares, are what keep him in the chair. The US$1.49 billion for Atome came from Grab’s account, not from his.
Footnotes
-
SoftBank held 18.6 percent of Grab’s stock at listing and just 7.6 percent of the vote, less say over a decision like Atome than the loan Tan’s mother made toward MyTeksi back in 2012. ↩