Alibaba: The Lawsuit Is Not the Point — The Business Reset Is

A securities class action deadline of October 5 has circulated through investor inboxes for Alibaba, with law firms asking shareholders who bought between June 26, 2025 and June 24, 2026 to step forward. The complaint alleges management misled investors about two things: whether Alibaba fell under the National Defense Authorization Act's definition of a "Chinese military company," and whether the company's AI development involved unauthorized use of competitors' models.

The lawsuit is a symptom, not the diagnosis. What actually moved Alibaba from a high near $173.68 on October 9, 2025 to $108 today was a cluster of events in June 2026 that forced a reckoning — and an operating picture that tells a more complicated story than either the lawsuit or the rally that preceded it.

What happened in June

Two disclosures in one month compressed the stock by roughly 25%. On June 8, the Pentagon added Alibaba to its Section 1260H list, citing ties to the Ministry of Industry and Information Technology. The stock fell 3.9% over the following two sessions. On June 24, Bloomberg reported that Anthropic had accused Alibaba of conducting the largest known "distillation attack" on its Claude AI model — roughly 25,000 fraudulent accounts generating 28.8 million exchanges between April and June to extract proprietary AI capabilities. Shares dropped another 4.7% the next day.

These events triggered the class action. But the legal filing tells you little about whether Alibaba is worth owning. The operating numbers do.

The earnings that don't match the narrative

Alibaba reported its fiscal 2027 first quarter on August 20. Revenue came in at roughly 269 billion yuan, up 9% year over year and slightly above the 268 billion yuan consensus. The headline looked fine. Then the profitability picture appeared.

Earnings per share of 8.52 yuan missed the 10.72 yuan estimate. Adjusted EBITA fell 30% to 27.3 billion yuan, and the adjusted margin collapsed from 16% to 10%. Over the trailing twelve months, the operating margin sits at roughly 5%, and free cash flow has swung sharply negative — down 177% year over year to minus 6.9 billion yuan.

The company is burning through cash building AI infrastructure. And here's the tension: the AI business is genuinely growing. AI cloud and compute services revenue climbed 45% to 48.4 billion yuan. AI-related product revenue has grown triple-digit for 11 consecutive quarters. Alibaba says AI products now represent 30% of cloud's external revenue and are on track to cross 50% within a year.

But 45% growth in one business doesn't matter if it costs more than it earns and the rest of the company is stagnating. China e-commerce customer management revenue fell 7%. The company that used to print margin is now spending it on compute.

The DoD designation: reputational shock versus operating consequence

The Section 1260H list has real teeth, but they don't bite Alibaba's core business yet. Phase 1, effective June 30, bars the Pentagon from directly contracting with designated companies. Phase 2, starting June 30, 2027, extends to indirect supply chain goods and services.

For a company whose revenue comes from Chinese e-commerce, cloud services sold primarily in Asia, and logistics, the Pentagon procurement ban is not a meaningful operating hit. Alibaba doesn't sell to the U.S. defense industrial base in any material way.

The reputational damage, however, is real. The designation signals that Washington views Alibaba as embedded in China's military-civil fusion apparatus. Alibaba has filed a lawsuit against the Pentagon and denies the allegation. More consequential is the secondary risk: if the Treasury Department adds Alibaba to the NS-CMIC list, U.S. persons would be prohibited from buying or selling the stock. That hasn't happened. But the possibility alone changes the risk calculus for every American holding these shares.

The Anthropic accusation compounds the reputational risk on the other side. Whether the distillation allegations are ultimately proven or resolved, they raise a question that goes to the heart of Alibaba's AI story: if the company's fastest-growing business depends on methods that Western competitors classify as hostile, how far can that business expand before the backlash becomes operational — not just political?

The valuation bridge

This is where the number does the work. Alibaba trades at a forward P/E of roughly 11 and a price-to-sales ratio of 1.7. Market capitalization sits around $268 billion. The company holds 21 billion yuan in cash with net debt of minus 17.6 billion — a net cash position. Total enterprise value is 251 billion.

The multiple is cheap by historical standards and by any growth-company standard. If the AI investments begin to pay off and margins recover from the current 5% operating level, the multiple could re-rate quickly. If they don't, the 11x forward earnings multiple is what growth companies trade at when the growth stops.

The company is buying back shares — repurchasing roughly 4.1 million in early July alone under an ongoing program — and it still pays a dividend, though the yield is modest at roughly 1%. These are signals that management believes the stock is undervalued. That doesn't prove they're right, but it means someone with access to the internal numbers is spending real cash on the bet.

What changes the call

The bear case is simple and it's live: AI spending is destroying free cash flow faster than revenue growth can absorb it, the e-commerce core is weakening, and the geopolitical overhang could escalate from reputational to operational at any time. If the Treasury NS-CMIC listing materializes, U.S. investors would face the practical problem of being unable to sell. If AI profitability doesn't emerge within the next two quarters, the 30% decline in adjusted EBITA becomes structural rather than transitional.

The bull case requires one thing to prove true: that Alibaba's AI investments are building a durable, defensible business rather than an expensive race to parity. The 45% AI cloud growth, the proprietary chip deployment, and the Qwen model family are the evidence for that claim. But evidence of spending is not evidence of economics yet. The company needs to show that AI revenue growth is translating into margin expansion, not just revenue that sits above a growing cost line.

The next earnings report and any Treasury action on the NS-CMIC list are the proof points. Until one of those shifts, the stock sits in the space between a beaten-down multiple and an unproven pivot — cheap enough to watch, but the business reset has to earn its valuation before it earns the buy.