Alibaba is Mispriced
Disclaimer: I’m just one investor thinking out loud. This isn’t financial advice. Do your own work before buying or selling anything.
Holy fuck is this thing oversold…
As I mentioned over the last couple weeks, the cheaper BABA gets heading into earnings, the more bullish I get. And man have we have gotten a lot cheaper.
The stock is sitting in the low $130s with an RSI around 22. That is full-on capitulation territory. The 52-week range is $96 to $193 and we’re a hell of a lot closer to the bottom of that range than the top, sitting well below every relevant moving average after a 30%+ drawdown from the highs.
And I don’t think this sell-off has much to do with Alibaba the business.
Asia has been getting absolutely killed this week. The Hang Seng dropped to two-month lows as the Iran situation blew up. U.S. and Israeli strikes, missiles going back the other direction, the Strait of Hormuz in play, oil ripping higher, etc.
South Korea’s Kospi fell so hard it triggered a trading halt. Hong Kong financials, tech, consumer, everything got smoked. When stuff like this happens, correlations go to one and everything sells together. Margin calls and forced liquidations made it worse. BABA just happened to be in the blast radius.
On top of that, there’s been a wave of selling from mainland Chinese investors pulling out of Hong Kong-listed shares. Reportedly billions of dollars worth in outflows over the past week as frustration builds over tech companies spending heavily on AI without showing earnings growth yet.
So you’ve got geopolitical panic layered on top of an AI sentiment unwind layered on top of forced selling. It’s a mess, and it has very little to do with whether Alibaba is a good business at a good price.
The Qwen situation, and what people are getting wrong
There’s been a lot of talk about the Qwen leadership departures this week. Junyang Lin, the tech lead, stepped down. Yu Bowen, the head of post-training, also resigned. That’s three senior departures from the AI unit this year. This certainly added fuel to the fire as BABA dropped another 4-5% in Hong Kong on the news, and I’ve seen a lot of people jumping to the conclusion that Alibaba is about to go closed-source and replicate the Google/Gemini playbook.
I don’t think it’s that simple.
Google can afford to run Gemini as a closed moat because GCP is a mature business with massive enterprise penetration. They already have the customers. Alibaba Cloud is nowhere near that scale. They're the undisputed leader in China and Asia, but globally they're still competing for share against GCP, AWS, and Azure while domestically fending off Bytedance who are cutting prices aggressively to try to claw their way in.
For Alibaba, open source IS the customer acquisition strategy. You don’t just torch the most popular open-source model family on earth to copy a playbook from a company in a fundamentally different competitive position.
What’s more likely happening is that Alibaba is tightening the link between model development and cloud monetization. The path from free downloads to paying cloud customers gets more deliberate. That doesn’t mean open source goes away. The frontier “Max” models have always been closed. The smaller models have always been open. That has literally been the strategy since Qwen 2.
So the leadership shakeup is definitely worth watching, but reading this as “Alibaba is abandoning open source” misreads the competitive situation. They need developers building on their infrastructure more than they need API margin on frontier models right now. At least that’s how I view it…
And for whatever reason I think most people are looking at the wrong part of the AI story with Alibaba anyway.
Yeah, the Qwen models are impressive. Qwen 3.5 is competitive with frontier models globally, the consumer app went from basically nothing to over 200 million monthly active users in two months, and they’re the most downloaded open-source AI models on the planet. I’m glad they’re doing well. But I actually think large language models get commoditized over time. In fact, we’re already seeing it.
The moats in AI aren’t going to be in the models themselves, they’re in the infrastructure layer underneath.
Alibaba Cloud is the clear leader in China. They have the full-stack offering (compute, storage, networking, the AI platform layer on top) and they’re the only Chinese cloud provider that can credibly compete across the entire stack. The only real comparable competitor is Bytedance, which has been cutting prices to try to win share, but Alibaba already has the installed base, the enterprise relationships, and the developer ecosystem.
Cloud revenue continues to accelerate and AI-related cloud products have posted triple-digit growth for nine consecutive quarters. Eddie Wu said on the last earnings call that they’re rationing GPU access because they literally can’t deploy servers fast enough to meet demand. They committed $53 billion over three years to AI and cloud infrastructure, and I’ve heard rumors they plan to increase that number.
Every company in China racing to build AI applications needs compute. And when they need compute, most of them are going to Alibaba.
Valuation
At $133, BABA is trading at roughly ~12x LTM EV/EBITDA. You can actually call this thing “cheap” again.
On top of that, I think this company is underearning right now. Margins have been compressed by all the heavy AI and cloud investment spending, competitive pricing pressure from Bytedance, and the quick commerce losses which I believe have peaked.
I think we’ve hit the bottom on margins and they start to improve from here. As that capex starts converting into paying cloud workloads and the pricing war stabilizes, you should see gradual but meaningful margin expansion.