Key points
- Alibaba owns an approximately 36% stake in Kimi's parent company, Moonshot AI, for about $800 million, per Alibaba's own SEC filing.
- Alibaba also holds pieces of MiniMax, Zhipu AI, and Kling AI, though those exact percentages are not public.
- Alibaba's own Qwen3.8-Max model claims to rank second only to Fable 5, an unverified claim from Alibaba itself, not an independent benchmark.
- Two unrelated hits landed on the stock in three weeks: a 7.3% drop after Anthropic accused Alibaba of illicit Claude access, and a $600 million DOJ settlement.
- At a $270 billion market cap, BABA trades near 17x forward earnings, and the 40-analyst average target is about 60% above the current price.
I keep looking at Alibaba (BABA) and my brain splits into two different stocks. One side owns a piece of nearly every hot AI company coming out of China right now. The other side has taken two ugly, unrelated headlines in the last three weeks. Let's get into both.
The AI stake nobody talks enough about
Everyone is talking about Kimi K3 this week. Fewer people are talking about who owns a third of the company behind it. We wrote yesterday about Kimi K3 pushing chip stocks into a bear market. What matters here is the size of Alibaba's check. Per Alibaba's own SEC filing, the company put in about $800 million for an approximately 36% stake in Moonshot, the startup behind Kimi.
That investment, based on the filing's own numbers, valued Moonshot at roughly $2.2 billion at the time. Moonshot is now raising a new round that reportedly values it at above $30 billion, and it just told investors it wants a Hong Kong IPO within six months. Do the math on that. A 36% slice of a $30 billion company is worth over $10 billion, on an $800 million check. I don't care how casual you are about a stock. That is a real number.
Moonshot isn't the only chip Alibaba has on the table. It also holds a stake in MiniMax, another Chinese AI company that already went public in Hong Kong this January and doubled on its first day. It's an investor in Zhipu AI too, alongside Tencent and Ant Group. And it just joined a $2.8 billion funding round for Kling AI, Kuaishou's video AI arm, on July 2. Alibaba hasn't disclosed exact percentages for any of those three, so I'm not going to pretend I know them. What I do know is the pattern. Alibaba isn't betting on one AI company. It's holding pieces of basically the whole board.
If even one of these companies goes public at a real number, Alibaba's stake stops being a talking point and becomes a line item.
And Alibaba built its own contender too
Alibaba isn't only collecting stakes in other people's AI companies. Its own model, Qwen, just put out a new version called Qwen3.8-Max on July 19, two days after Kimi K3 rattled the market. Alibaba's Qwen team called it one of the most powerful models available today and said it ranks second only to Anthropic's Fable 5.
That ranking comes from Alibaba itself, with no benchmark scores attached and no independent test confirming where the model actually lands. On LMArena's own public leaderboard, Fable 5 sits at the top, and the previous Qwen version ranked well behind it. Treat "second only to Fable 5" as a marketing line for now, not a verified result.
Real or not, the model itself is a big swing. It runs 2.4 trillion parameters, it's the first Qwen model over a trillion parameters that can handle images, video, and documents alongside text, and Alibaba is pricing the preview at 10% of standard rates to get people using it. Between Qwen and the stakes in Moonshot, MiniMax, Zhipu, and Kling, Alibaba is playing this AI moment from every angle at once, building its own model while owning pieces of the ones beating it.
Two headlines that had nothing to do with AI
None of that upside stopped this stock from taking two real hits in the last three weeks, for reasons that have nothing to do with Kimi or Moonshot.
On June 24, the Financial Times reported that Anthropic had accused Alibaba of getting illicit access to Claude by creating fake accounts. BABA's American shares fell 7.3% over the next two trading sessions. It wasn't an isolated shot. We covered it here when China's own security database called Claude Code a backdoor back on July 8, and Alibaba banned it internally. This looks like a real, ongoing standoff between Alibaba and one of the biggest names in AI, and it has already moved the stock once.
Then on July 1, the Justice Department announced Alibaba agreed to pay $600 million in a non-prosecution deal, settling claims that its platforms let merchants sell illegal pharmaceuticals and pill presses to buyers in the US for years. The stock dropped almost 2% on that news alone.
Add it up and you get a stock that can take a real hit on a random Tuesday for reasons that have nothing to do with Kimi, Moonshot, or any of the AI upside.
The risk that never actually goes away
Then there's the risk sitting underneath every US listed Chinese stock, not just this one. Treasury Secretary Bessent has said delisting Chinese shares from US exchanges is still on the table. Roughly 40% of Alibaba's shareholder base is retail investors, which means if that risk ever became real, a lot of regular people would be trying to sell at the same time. It probably isn't happening tomorrow. But it's the kind of tail risk that never shows up on a normal chart, and it doesn't disappear just because Kimi had a good week.
So is it a buy
My honest read: Alibaba is one of the few ways to own a diversified slice of China's entire AI boom instead of betting on a single name. Not many stocks let you do that. But this stock has now shown me twice in three weeks that it can eat a 2 to 7% hit from something that has nothing to do with the AI story I actually want to own it for.
I'd rather hold a small position here and actually sleep at night than load up and hope the next headline is another Anthropic accusation instead of something bigger.
One number keeps pulling me back in. Alibaba's total market cap is $270 billion, and the average target from 40 analysts is $190 per share, about 60% above where the stock trades today. That gap is partly the AI story and partly because the stock trades at roughly 14 times forward earnings, cheap next to US tech names running 25 to 35 times. Wall Street's own read on this is basically binary. If the cloud and AI bets pay off, that gap closes fast. If they don't, the stock probably stays cheap a while longer.
If Moonshot's IPO actually happens in the next six months, that's a real catalyst worth watching closely. Until then, I'm treating this as a stock to size carefully, not one to chase.
Sources
- Alibaba's approximately 36% Moonshot AI stake and $800 million investment: Alibaba Group Holding Limited, Form 20-F, fiscal year 2026, SEC filing.
- Anthropic's fake account access accusation against Alibaba: Financial Times, June 24, 2026 reporting, and the stock's two session decline.
- DOJ non prosecution agreement and $600 million settlement: US Department of Justice announcement, July 1, 2026.
- Moonshot AI funding round and Hong Kong IPO plans: Bloomberg, July 19, 2026.
- MiniMax and Zhipu AI stake and IPO context: company disclosures and financial press coverage, July 2026.
- Alibaba market cap, forward P/E, and 40-analyst average price target: stockanalysis.com, as of July 20, 2026.
This is my own opinion and general commentary, not investment advice. I do not currently hold a position in Alibaba (BABA). Always do your own research and consider talking to a licensed financial professional before making any investment decision.
