Stock Pick #5 of my 2026 series — Consumer Discretionary
Last week I wrote about WeRide, my fourth stock pick under Industrials. This week I’m doing something harder — re-evaluating one of my biggest holdings in what has been a pretty uncertain market.
Alibaba has been sitting at ~16% of my portfolio for a while now. Every quarter I told myself the next one would be better. The cloud story would play out. The AI investment would pay off. Then I looked at the last two quarters properly and the numbers told a different story. A stock this volatile, with a strategy this unclear, has no business taking up 16% of my portfolio. I’ve decided to start trimming the position.
The BABA Problem I’ve Been Sitting On
It rallied hard in early 2025 — up 75%, peaking around USD 192. I was sitting on a decent gain through most of 2025. Then it drifted back, and coming into 2026 I’m now sitting on a loss.
The more I dug into the numbers, the harder it became to justify holding it at this size:
Alibaba — Six Quarter Snapshot
| Metric | Q1 FY25 Jun’24 | Q2 FY25 Sep’24 | Q3 FY25 Dec’24 | Q4 FY25 Mar’25 | Q1 FY26 Jun’25 | Q2 FY26 Sep’25 |
|---|---|---|---|---|---|---|
| Revenue (RMB bn) | 243.2 | 236.5 | 280.2 | 236.2 | 247.7 | 247.8 |
| Revenue Growth (YoY) | +4% | +5% | +8% | +7% | +10% | +15% |
| Profit Margin | 19% | 17% | 20% | 14% | 16% | 4% |
| Earnings per ADS (RMB) | 16.44 | 15.06 | 21.39 | 12.52 | 14.75 | 4.36 |
| Free Cash Flow (RMB bn) | +17.4 | +13.7 | +39.0 | +3.7 | -18.8 | -21.8 |
For most of 2025, the numbers were decent. Margins healthy, cash flowing in, earnings growing. That’s what drove the stock up 75%.
Then something broke. In the last two quarters, profit margins collapsed from 20% to just 4%. Earnings dropped 71% year-on-year. And free cash flow — the actual cash coming into the business — flipped from +RMB 39 billion to -RMB 22 billion. A company sitting on RMB 574 billion in cash is now burning through it.
What’s strange is that revenue is still growing — actually accelerating. So where is the money going?
Four places.
The core e-commerce business is under pressure. Taobao and Tmall are still the biggest marketplace in China, but they’re losing ground to PDD and ByteDance’s Douyin. To keep revenue growing, Alibaba has been raising fees on merchants — not because more people are buying, but to squeeze more from each transaction. That works until merchants start leaving for cheaper platforms. It’s a short-term fix with a long-term cost.
Then there’s the food delivery war. Earlier in 2025, Alibaba launched an RMB 50 billion subsidy campaign to take on Meituan and JD.com — and it hurt badly. All three platforms burned through over RMB 100 billion combined. Regulators eventually stepped in and forced a truce by mid-2025. The bleeding has stopped, but the damage is done. More importantly, it raises a bigger question: was food delivery ever Alibaba’s fight to pick?
AI is both the opportunity and the cost. The Qwen LLM models are genuinely competitive — Qwen3.5, released in February 2026, benchmarks well against leading US and European models. Cloud is growing 34% and AI-related revenue has posted triple-digit growth for nine straight quarters. But building frontier AI models is expensive, and US export restrictions mean Alibaba can’t get Nvidia’s best chips. They’re spending more to achieve less.
And then there’s just the reality of holding any Chinese stock listed in the US. Policy changes, export restrictions, political noise — any of it can move the stock on any given day regardless of what the business is actually doing. That’s the deal you accept going in.
The Qwen models are good enough that Alibaba is powering the 2026 Milano Cortina Winter Olympics with them. And to be fair, I haven’t tried Qwen myself — so I’m not dismissing the technology.
But here’s my problem. What exactly is Alibaba’s main business now? E-commerce? Cloud? Food delivery? AI? Eddie Wu has been CEO since September 2023 and I genuinely cannot answer that question. I’ve watched too many companies try to be everything at once and lose sight of what made them good in the first place. Good technology doesn’t matter if no one can tell you what the company is trying to be.
I used to compare BABA to Amazon — cloud as the future profit engine, e-commerce as the foundation. It’s not a wrong comparison on paper. But look at the margin gap:
| AWS (Amazon) | Alibaba Cloud | |
|---|---|---|
| Revenue contribution | ~15% of total | ~12% of total |
| YoY revenue growth | ~18% | ~34% |
| EBITA margin | ~37% | ~9% |
Alibaba Cloud is growing twice as fast but earning less than a quarter of the margin. AWS doesn’t just grow — it funds everything else Amazon does. Alibaba Cloud isn’t there yet. It’s still in heavy investment mode, chip restrictions are making it more expensive to compete, and meanwhile the e-commerce moat is quietly eroding and cash was burned on a food delivery war that shouldn’t have been fought.
Then on March 4, Lin Junyang — the technical lead of the Qwen division and its most visible public voice globally — announced he was stepping down, with no explanation given. Two other senior Qwen researchers left the same day. He’s now the third senior Qwen executive to depart in 2026. What matters is whether the strategy is clear enough for whoever stays to execute it. Right now I’m not sure it is.
Why Not PDD?
Before I get to Sea Limited (NYSE: SE), I should explain why I didn’t go to PDD Holdings (NYSE: PDD) instead.
| Alibaba (BABA) | PDD Holdings (PDD) | Sea Limited (SE) | |
|---|---|---|---|
| Revenue Growth | ~9% | ~12% | ~36% |
| Fwd P/E | ~20x | ~8.5x | ~20x |
PDD looks cheap at 8.5x forward earnings. But that number reflects a real problem. Temu was built on a tax rule that let cheap Chinese packages enter duty-free — the US closed that loophole in May 2025, and Europe is following from July 2026. Since then, Temu’s US daily active users dropped over 50% and ad spend was cut 95% year-on-year in May 2025. PDD’s response is to rebuild Temu around local warehouses, EU-based merchants, and compliance infrastructure — a fundamentally different business from the one that grew 59% in 2024. That kind of transition takes years and the right execution. I’m not convinced.
Sea Limited is where I landed.
Why Sea Limited
I actually have history with Sea. I first bought in late 2022 and spent the next year trading in and out — buying around USD 65, selling around USD 72, then rotating into PDD and eventually BABA. It worked well enough when markets were calmer and momentum was more predictable.
Coming back to Sea now, the business I’m looking at today isn’t the one I was trading in 2022. Back then, Garena — the gaming arm — was still carrying the company. Now the picture is completely different, and what I like most about it is how deliberate the whole thing looks.
Garena generated the early profits. Those funded Shopee’s expansion across Southeast Asia. Shopee’s transaction data then gave Monee — their fintech arm — the ability to lend money to people that traditional banks simply can’t assess. Each business built on the one before it, on purpose. Today Shopee is the dominant platform in Southeast Asia — USD 127 billion in GMV in 2025, the largest in the region. The strategy was clear from the start and it’s playing out.
Sea Limited — Five Quarter Snapshot
| Metric | Q4 2024 | Q1 2025 | Q2 2025 | Q3 2025 | Q4 2025 |
|---|---|---|---|---|---|
| Revenue (USD bn) | 5.0 | 4.8 | 5.3 | 6.0 | 6.9 |
| Revenue Growth (YoY) | +37% | +30% | +38% | +38% | +38% |
| Net Income (USD mn) | 238 | 411 | 414 | 375 | 411 |
| Net Income Margin | 4.8% | 8.6% | 7.8% | 6.3% | 6.0% |
| Adj. EBITDA (USD mn) | 591 | 947 | 829 | 874 | 787 |
| Adj. EBITDA Margin | 11.8% | 19.7% | 15.6% | 14.6% | 11.4% |
| — Revenue Mix — | |||||
| Shopee (% of revenue) | 64% | 72% | 72% | 63% | 62% |
| Monee (% of revenue) | 15% | 16% | 17% | 17% | 16% |
| Garena (% of revenue) | 10% | 10% | 11% | 11% | 10% |
| — Segment Growth (YoY) — | |||||
| Shopee | +41% | +29% | +34% | +35% | +38% |
| Monee | +55% | +58% | +59% | +61% | +54% |
| Garena | +35% | +8% | +28% | +31% | +35% |
Remaining ~10% is “Sales of Goods” (Shopee 1P retail). Core 3 segments account for ~90% of revenue.
But if I had to pick one segment, Monee is the one I’m most excited about. Growing 54-61% every quarter, loan book at USD 9.2 billion, bad loans at just 1%. That’s a fintech business quietly becoming a serious profit engine.
In February 2026, Sea and Google announced a partnership to build AI shopping tools for Shopee. Sea is headquartered in Singapore, and its founder Forrest Li has been a naturalised Singaporean citizen since the late 2000s — no chip restrictions, no Pentagon blacklist risk, no Beijing regulatory overhang. This isn’t Alibaba or ByteDance.
I’ve started building a position in SE. I’m treating this as a 2-3 year hold, not a trade.
What Could Go Wrong
The thing I watch most closely is margins. Group EBITDA is growing but the trajectory is declining — from 19.7% in Q1 2025 to 11.4% by Q4. Management is investing heavily in logistics, fulfillment, and AI — including the Google partnership on Shopee’s shopping tools — but these are still costs, not returns yet. Sea is where Amazon was before it started harvesting its infrastructure investments. On the Q4 2025 earnings call, management guided for EBITDA no lower than 2025 in absolute dollars. That’s a floor, not a target.
Monee is the other risk. The credit loss provision jumped 67% in Q4 2025, which is what really spooked the market on March 3, 2026. The loan book grew 80% to USD 9.2 billion, and in Q4 2025 alone Monee added 5.8 million first-time borrowers under an “all-can-apply” approach. That’s aggressive expansion by any standard. The difference from other ecosystem lenders in the region — Grab Financial, GoPay — is the depth of Shopee’s transaction data. Grab’s NPL runs at 1.8% on a USD 566 million loan book. Monee’s is 1.1% on USD 9.2 billion. Short-duration loans mean credit limits can be tightened quickly if quality deteriorates. But off-Shopee lending has surged over 300% and now makes up 15% of the book — loans made without that behavioral data advantage. The question is whether that holds when the macro turns and the model gets its first real stress test.
The competitive risk worth watching is TikTok Shop. Its SEA GMV doubled to USD 45.6 billion in 2025 and it’s closing the gap through social commerce and live streaming — formats where Shopee is still catching up. Lazada — Alibaba’s vehicle in the region — has dropped to under 8% market share and is no longer a serious threat. TikTok is. Shopee is responding with its own live commerce features, but this is a real fight.
Garena’s dependency on Free Fire is a real concentration risk. But Free Fire has posted over 30% bookings growth for two consecutive years, and Garena’s EBITDA grew 25% in Q4 2025. The single-game risk is worth watching — it’s just not what the numbers are saying right now.
Where I Stand
On BABA: I’m going to start trimming, not exiting. The cloud business is real, the AI investment is real, and there’s still a growth story here if Eddie Wu can bring strategic clarity. But at 16% of my portfolio, the position is too large for the uncertainty it carries. The next quarterly results will be the first real test of whether that strategic clarity is starting to show up in the numbers. That’s what I’ll be watching.
On SE: Currently around 4% of my portfolio. When fear is doing the selling and the fundamentals haven’t changed, that’s usually where the opportunity sits.
Disclaimer: This is not financial advice. I hold a position in Alibaba (BABA) and have initiated a position in Sea Limited (SE). All investors should conduct their own research and consider their own risk tolerance before making investment decisions.