Alibaba's Multiple Went From Deep Value to Peak AI Optimism — Then Nearly All the Way Back

Introduction

On October 2, 2025, Alibaba Group Holding (BABA) traded above $192 a share, its highest print in more than four years, on a story that practically wrote itself: a Chinese tech giant turning its cloud division into an AI powerhouse just as the rest of the market was falling in love with anything adjacent to large language models. Eight months later, on June 26, 2026, the same stock touched $91.99 — a 16-month low, down roughly 33% year-to-date — after Anthropic told the White House that Alibaba's Qwen research lab had run what it called the largest known distillation attack against Claude. By August 17, 2026, two trading days before Alibaba's fiscal first-quarter earnings call, the stock had clawed back to $124.71, still nowhere near its October high but well off its June trough.

That is not a normal glide path from undervalued to overvalued. It is a full round trip — deep value to peak optimism to crisis and partial recovery — compressed into less than two years, and it happened almost entirely on the back of one storyline: artificial intelligence. Untangling which parts of that swing reflect Alibaba's business actually changing, and which parts reflect the market's mood about Chinese AI shifting underneath it, is the point of this piece.

Key Takeaways

• Alibaba's forward P/E ran from roughly 11x-12x in early 2024 to an AI-euphoria peak near 33x-35x around October 2025-January 2026, collapsed to roughly 14x-15x by mid-2026, and has since rebuilt to 18.96x as of August 17, 2026 — still well below its October 2025 peak (data via YCharts, cited in Seeking Alpha's Aug. 17, 2026 earnings preview, "Alibaba Remains Too Cheap To Ignore").
• The stock's June 2026 crash to a 16-month low was triggered by a specific event, not a slow bleed: Anthropic notified the White House on June 24, 2026 that Alibaba's Qwen lab allegedly ran roughly 28.8 million interactions with Claude through about 25,000 fraudulent accounts between April and June 2026 to extract its capabilities (TheStreet, June 2026).
• Alibaba has missed Wall Street's revenue estimate in 7 of the past 12 quarters and its EPS estimate in 7 of 12, including five straight quarters of EPS misses — yet its fiscal Q4 2026 (March quarter) actual EPS of $0.09 against a $1.12 estimate, a roughly 92% miss, reflected heavy AI and cloud investment rather than a weakening core business (TipRanks earnings data, reported May 13, 2026).
• The recovery since June has been built on real momentum: Qwen's open-weight models crossed 3 billion global downloads in six months, overtaking Meta and Google on Hugging Face's tracker; Alibaba Cloud holds an estimated 40.1% share of China's full-stack AI cloud market; and Apple has partnered with Alibaba on a custom AI model to finally bring Apple Intelligence to China (Bloomberg, Aug. 15, 2026; Reuters via Seeking Alpha, Aug. 2026).
• Real risk remains attached to that recovery: a Rosen Law Firm securities class action filed August 4, 2026 over the Anthropic allegation, an Axios report that the Trump administration is weighing a ban on foreign open-source AI models specifically after Qwen 3.8's launch, and an EU parcel fee squeezing AliExpress — all while Alibaba heads into an August 20 earnings call it has beaten only rarely over the past three years.

Four Windows in Twenty Months

Chart Alibaba's forward price-to-earnings multiple from early 2024 through today and the shape is closer to a spike than a trend: a long, quiet baseline, a sharp climb into euphoria, an even sharper collapse, and a partial, still-incomplete rebuild.

• January 2024 (The Quiet Baseline): Forward P/E sat around 11x-12x, a level consistent with a company still working through years of regulatory overhang from Beijing's tech crackdown and slow e-commerce growth. Genuinely the cheapest window of the four, even cheaper on a forward-multiple basis than the "crisis" trough that would come 18 months later.
• October 2025-January 2026 (Peak AI Optimism): The multiple spiked to roughly 33x-35x as the stock hit its 52-week (and multi-year) high of $192.67 on October 2, 2025 — a re-rating built on Alibaba's AI infrastructure pledge and the broader 2025 China-AI rally.
• Mid-2026 (The Reckoning): By the time the stock bottomed at $91.99 on June 26, 2026, the forward multiple had fallen back to roughly 14x-15x — still above the 2024 baseline, because Wall Street's earnings estimates were being cut nearly as fast as the price, but unmistakably a return to skepticism after a year of euphoria.
• August 17, 2026 (The Rebuild): The forward multiple has recovered to 18.96x, roughly half its October 2025 peak, on a wave of concrete AI and cloud news — but still pricing in real doubt about whether Alibaba's AI narrative and its actual profitability are moving in the same direction.

Notice that the cheapest bar on the chart isn't the "crisis" window everyone remembers — it's the quiet one from early 2024, before anyone was talking about Alibaba as an AI story at all. That's the detail worth sitting with before assuming June 2026's crash pushed the stock back to genuine rock-bottom pricing. It didn't, quite — because analysts were cutting their earnings estimates almost as fast as the price was falling, the "E" in the P/E ratio was shrinking right alongside the "P."

October 2025: A Rally Built on Infrastructure Spending and a Comeback Narrative

Alibaba's run to $192.67 wasn't built on a single earnings beat — the company's own track record argues against betting on beats at all, given it has missed Wall Street's revenue estimate in 7 of the last 12 quarters. Instead, the 2025 rally was a story stock in the purest sense: CEO Eddie Wu's pledge, announced in early 2025, to spend roughly 380 billion yuan on AI infrastructure over three years, combined with Jack Ma's re-emergence in Alibaba's public narrative, fed a rally that CNBC at the time pegged at roughly $100 billion in added market value, tied explicitly to "AI, Jack Ma return" (CNBC, March 2025). Layered on top through the back half of 2025 was a broader China-AI re-rating — the same DeepSeek-adjacent wave that convinced global investors Chinese labs could match Western AI capability at a fraction of the cost — which helped push China's tech-heavy STAR 50 index up roughly 65% in the first half of 2026 even as bubble warnings from firms like Morgan Stanley and UBS started circulating by mid-year (South China Morning Post, 2026).

A forward multiple in the mid-30s, on a company that misses its own quarterly estimates more often than it beats them, is a bet that AI infrastructure spending converts into durable, high-margin revenue years before it shows up in any given quarter's income statement. That is precisely the kind of bet that tends to look safest right before it stops paying off.

June 2026: How a Single Letter to the White House Erased Eight Months of Gains

Unlike the slow, macro-driven multiple compressions that show up in most valuation histories, Alibaba's 2026 reckoning had an identifiable trigger date. On June 24, 2026, Anthropic notified White House officials and members of Congress that it had evidence Alibaba's Qwen research lab ran what several outlets described as the largest known distillation attack against Claude — an alleged campaign of roughly 28.8 million interactions with Claude models through approximately 25,000 fraudulent accounts between April and June 2026, aimed at extracting Claude's capabilities to train Qwen (TheStreet, June 2026). The timing lined up with an April 24, 2026 White House memo that had already warned about industrial-scale foreign distillation of U.S. AI models, giving the accusation an immediate policy audience.

The market's reaction was fast and severe: BABA shares fell as much as 4.9% in a single session, extending the stock's 2026 year-to-date decline to roughly 33% and sending it to a 16-month low. The stock bottomed at $91.99 on June 26, 2026 (stockanalysis.com, accessed Aug. 18, 2026). High-profile investors amplified the move — both Cathie Wood's ARK funds and Michael Burry disclosed trimming or exiting BABA positions around the same period, adding a visible bearish signal on top of the news itself.

The crash landed on top of an already-fragile fundamental picture. Alibaba's fiscal Q4 2026 (March quarter) results, reported May 13, 2026, showed actual EPS of $0.09 against a $1.12 estimate — a roughly 92% miss against the quarter's own consensus, its fifth consecutive quarter of EPS disappointment. And the Chinese consumer backdrop wasn't helping: growth in China's "618" mid-year shopping festival, a bellwether for domestic e-commerce demand, slowed sharply to just 4% year-over-year in June 2026, down from 15.2% the year before, according to Syntun data reported by CNBC — evidence that the consumption side of Alibaba's business, not just its AI ambitions, was losing momentum at the same moment the Anthropic story broke (CNBC, June 23, 2026).

By early August 2026, the fallout had moved from the market to the courtroom: Rosen Law Firm filed a securities class action on August 4, 2026, covering a class period from June 26, 2025 through June 24, 2026, alleging Alibaba failed to adequately disclose both potential "Chinese military company" designation risk under U.S. defense-authorization rules and the ongoing nature of the distillation-attack exposure (BusinessWire, Aug. 4, 2026). That litigation remains unresolved heading into earnings.

July-August 2026: The Rebuild, Built on Real News This Time

The recovery from $91.99 to $124.71 wasn't sentiment alone — it tracked a genuine run of AI and cloud announcements. Reuters reported in mid-August 2026 that Apple had partnered with Alibaba to build a custom large language model for the Chinese market, positioning Alibaba's technology inside Apple Intelligence as it finally clears Chinese regulatory approval nearly 22 months after its original global launch — with Apple reportedly the first foreign company authorized by Beijing to deploy a proprietary AI model domestically, and Baidu handling the search-integration piece (Reuters via Seeking Alpha, Aug. 2026; 9to5Mac, Aug. 14, 2026).

On the infrastructure side, Alibaba Cloud is estimated to hold a 40.1% share of China's full-stack AI cloud market — ahead of Baidu, ByteDance's Volcano Engine, and SenseTime combined, according to a Frost & Sullivan report cited across multiple outlets — and in early August 2026, Bloomberg reported Alibaba had agreed to supply Moonshot AI with roughly 20,000 Nvidia chips to train Moonshot's Kimi models, a deal that sent Alibaba's Hong Kong-listed shares up as much as 7% on the news (Bloomberg via Yahoo Finance, early Aug. 2026). And on adoption, Alibaba's open-weight Qwen models crossed 3 billion global downloads over the trailing six months as of mid-August 2026, overtaking Meta's 227 million and Google's 418 million on Hugging Face's "State of Open Models" tracker — with more than 460 open-sourced models and over 300,000 community-built derivatives (Bloomberg, Aug. 15, 2026). Alibaba is reportedly planning to monetize that adoption directly, asking major commercial users of the next open-source Qwen release for a share of the revenue they generate from it — a meaningful shift from giving the models away for free.

Wall Street's read on all of this is genuinely split, which is itself useful information. Bank of America cut its BABA price target from HK$176 to HK$168 on April 12, 2026 — while keeping a Buy rating — specifically citing rising AI investment and capital-expenditure pressure on margins (GuruFocus, Apr. 12, 2026). Susquehanna, by contrast, set a $185 price target on May 15, 2026, squarely in the more bullish camp. The broader analyst consensus remains a Strong Buy — roughly 38-40 analysts covering the stock, only one at Sell — with average 12-month price targets clustering between roughly $187 and $192, implying 50%-plus upside from the current price (Benzinga, Aug. 2026). At the current price, Seeking Alpha's own data shows the stock trading around 10.55x 2029 earnings estimates — a multiyear view under which even today's rebuilt 18.96x forward multiple still looks inexpensive, if the estimates hold.

That's a real "if." Two live risks sit directly on top of the recovery narrative. First, Axios reported on July 20, 2026 that the Trump administration is reviving efforts toward a de facto ban on foreign open-source AI models, a push triggered specifically by the launches of Moonshot's Kimi K3 and Alibaba's Qwen 3.8 — with options reportedly including Commerce Department Entity List additions, an executive order targeting foreign open-source models, and federal procurement bans, even as David Sacks, chair of the President's Council of Advisors on Science and Technology, has publicly warned such a ban could stifle U.S. AI competitiveness rather than protect it (Axios, Jul. 20, 2026). A separate, distinct Axios report from July 28, 2026 covered Chinese hardware restrictions tied to the AI race — a different lever, not to be confused with the open-source-model ban under discussion. Second, on the consumer side, the European Union's new €3 flat customs duty on parcels under €150 from non-EU sellers, effective July 1, 2026 and replacing the prior duty-free threshold entirely, is squeezing the price competitiveness of Alibaba's AliExpress marketplace in one of its largest cross-border markets, with a typical small multi-item order potentially seeing total costs rise 30%-50% once duties, VAT, and handling fees layer on (Cybernews, Jul. 2026).

What Thursday's Print Actually Needs to Show

Alibaba reports its fiscal first-quarter 2027 (June quarter) results before the market opens on Thursday, August 20, 2026 — and as of this writing, that report has not yet happened. Consensus calls for revenue of $39.80 billion, implying 14.58% year-over-year growth, an acceleration from the 11% comparable growth Alibaba posted in its March quarter. EPS consensus sits at $1.59, a projected 23.14% decline year-over-year — a smaller drop than the 95% year-over-year EPS decline the company posted in its March quarter, but still a clear signal that AI and cloud investment will keep weighing on near-term profitability (figures per Seeking Alpha, "Alibaba Remains Too Cheap To Ignore," Aug. 17, 2026). Worth noting: other data providers show meaningfully different EPS estimates for the same quarter, ranging as low as $1.49 and as high as $1.94, which on its own tells you how unsettled sentiment is heading into the print.

History argues against assuming an easy beat. Alibaba has missed Wall Street's revenue estimate in 7 of its past 12 quarters, including 6 of the last 8, and has missed on EPS in 7 of 12, with five consecutive quarters of EPS disappointment through the March 2026 report. Analyst revisions in the three months leading into this print have been close to a coin flip — four upward and five downward revisions on revenue, four up and four down on EPS — which is itself a fair summary of where sentiment sits: genuinely divided, not quietly bullish or bearish.

The Part a Forward Multiple Can't Settle on Its Own

Line up Alibaba's four windows and the lesson isn't that the stock is now cheap or expensive in some absolute sense — it's that almost none of the swings in this multiple have tracked Alibaba's actual trailing profitability in any simple way. The 2025 rally ran on an infrastructure-spending pledge and a comeback narrative, not on beaten estimates. The 2026 crash ran on a single geopolitical and legal allegation landing on top of an already-soft consumer quarter, not on a change in Alibaba's underlying AI capability. And the current rebuild is being priced on genuinely verifiable adoption numbers — 3 billion Qwen downloads, a 40.1% China AI-cloud share, an Apple partnership — sitting directly next to unresolved legal exposure and a live U.S. policy threat that could specifically target the open-source models driving that adoption.

None of that makes 18.96x an obviously wrong price for Thursday's print. But it does mean the multiple is doing something closer to pricing a bet on which story wins — AI-driven growth or AI-driven geopolitical risk — than pricing what Alibaba has actually earned lately. A forward-earnings framework exists precisely because a multiple built on what a company is expected to make over the next twelve months tells you something different than one reacting to a single quarter's shock, good or bad — and Alibaba's last twenty months are as clean an illustration of that gap as any company in this series has offered so far. More on why we build valuation around forward estimates rather than trailing results.

Conclusion

Alibaba heading into its August 20 earnings call is a company being priced on a future that hasn't been confirmed yet — real AI adoption metrics, a landmark Apple partnership, and a dominant China cloud position, all counterbalanced by an unresolved securities lawsuit, a live U.S. policy threat aimed squarely at the open-source models it's celebrating, and a quarterly track record that has disappointed more often than it has delivered. The forward multiple has already told half the story by round-tripping from 11x to 34x to 14x to 19x in twenty months; what Thursday's actual numbers do next will decide whether the current 18.96x settles in as a fair discount for real risk, or turns out to be either too cautious or too generous, the way the last three re-ratings all eventually did.

---

This article discusses Alibaba's historical stock performance and the macroeconomic and geopolitical conditions surrounding it as of August 18, 2026, for informational purposes only. It is not investment advice or a recommendation to buy or sell any security. Alibaba's fiscal first-quarter 2027 (June quarter) results, scheduled for release before market open on August 20, 2026, had not yet been reported at the time of writing, and the figures, estimates, and analyst commentary referenced here are subject to revision.