Sea Limited SE

Key data

The business

Sea Limited runs three businesses off one Southeast Asian customer. Shopee is the marketplace, the largest online shopping platform across Indonesia, Vietnam, Thailand and five more markets, plus a fast-growing Brazil operation; it earns by taking a cut of every sale and selling ads to the sellers on it. Garena is the games arm, built on Free Fire, the mobile shooter that funded the whole company before it earned a dollar anywhere else; its bookings rise and fall with that one franchise, so it is the cyclical piece and it has whipsawed before, through a slump and an India ban. SeaMoney is the newest and highest-margin engine: it lends small sums to Shopee's shoppers and merchants and collects interest. The flywheel is real, a shopper on a Jakarta doorstep opening a Shopee package paid for with a SeaMoney loan, and it is also the whole risk, because the same customer carries all three engines at once. The moat is scale in a region global platforms have found hard to serve, held together by local logistics and a payments network few rivals can rebuild from scratch.

The numbers

Sea earns about $27.7B over the trailing year at a 44% gross margin, converts roughly 6% of that to net profit and more to cash, and carries almost no net debt. Earnings per share rose again in the June quarter and cleared the estimate, the second straight beat after two quarters late in 2025 that fell well short. That stumble matters. It is the tell that this model's profit still swings quarter to quarter on marketing and credit choices, not a smooth compounder yet.

The recovery is genuine but young. Analysts model the top line growing about 20% a year to roughly $45B by 2028, with per-share earnings nearly doubling to near $6, and that doubling is what the forward multiple rests on.

Here is the part that should stop you. Sea has de-rated far harder than its growth has slowed: the multiple has fallen more than two-thirds since it first turned a profit, and today the stock costs less per dollar of earnings than in any profitable year of its life.

Cheap or trap is the question. On repeatable earnings it reads as a market that repriced the late-2025 miss and has not marked the recovery back up. What this memo believes that the tape does not: the two recent beats are the start of real margin leverage, not a bounce, and the print that settles it is SeaMoney's loan losses set against Shopee's take rate next quarter.

Management

Sea is founder-run and founder-controlled: Forrest Li built it and holds the votes, with co-founder Gang Ye alongside. That cuts both ways. Over the last year insiders sold about $40M across roughly a hundred transactions and bought nothing, the largest a pair of late-August sales by Li Xiaodong and one by Ye Gang. The feed does not carry 10b5-1 status, so plan status is not disclosed; read the cluster as ordinary selling into a rally, not a signal, but note there was no answering purchase while the stock sat at barely half its high. What the pay package rewards, and the buyback record, were not pulled this run.

How it fails or surprises you

SeaMoney credit turns (downside). The loan book's delinquency and loss trend were not pulled this run, and that is the number that matters most: unsecured lending in Indonesia and Brazil books profit now and losses later. If consumer credit in either market rolls over, the highest-margin engine reverses first and fastest, and the earnings climb breaks with it. Watch the reported loss rate.

Shopee's take rate gets competed away (downside). TikTok Shop and Temu are spending hard to win the same baskets. Sea can defend share or defend margin, rarely both at once, and the late-2025 profit miss is what that fight looks like in the numbers. A falling monetization rate on flat GMV would show it inside a quarter or two.

Margin leverage compounds (right tail). If SeaMoney and Shopee ads keep dropping to the bottom line on 20% top-line growth, earnings roughly double into 2028 and today's 41-times multiple falls to the high teens on that year's profit. The market is pricing the stumble, not the leverage. A third clean beat with rising margin is the first proof, and few own the stock for it.

Closing thoughts

Next quarter's results will show whether the recovery is real. Two numbers settle it: SeaMoney's loan-loss rate and Shopee's take rate. If losses stay contained while monetization holds, the recovery is real and the discount to its own history is the opportunity; if either cracks, the late-2025 miss was the preview, not the exception. The fatter tail, on two straight beats and a multiple already cut by two-thirds, points up, but the downside is a genuine credit cycle in frontier markets, a real way to lose money, not a wobble.

The bet is still that Shopee keeps winning more of Southeast Asia's online shopping, across Indonesia, Vietnam, Brazil and five other countries, and that its lending arm keeps handing small loans to those same shoppers without too many failing to pay. It breaks if the people buying the packages stop repaying the loans. The pair to watch is the loan-loss rate against the take rate; hold the view only while both hold.