Corporate reorgs are never easy. 

When Alibaba proposed its “1+6+n” corporate reorg in March, breaking up the giant into one single holding company with six loosely held, somewhat independent businesses, plus a cohort of other assets (the “n”), it was the mother of all reorgs. At the time, I gave both a charitable and a cynical interpretation of what this undertaking could mean for Alibaba’s future. Since I’m naturally optimistic, I ended up taking the charitable side of the equation.  

So far, it’s been a hot mess and not going well, proving that the cynical version is more prescient. Since the initial reorg announcement, the status of Daniel Zhang, its CEO before the reorg, went from leading both the holding company and AliCloud, to just being CEO of AliCloud, to resigning entirely and just making venture investments off the corporate balance sheet. Joe Tsai became the group’s chairman. Eddie Wu became the group’s CEO. Wu also took over AliCloud from Zhang, but only on an acting basis. Freshippo, its groceries retail chain and one of the “n’s”, was gearing up for an IPO, but that has been put on hold. Cainiao, its domestic and international logistics unit, has filed its paperwork for a Hong Kong IPO, but the listing hasn’t been priced. Given this year’s difficult IPO market overall (see Arm, Instacart, Klaviyo’s share prices), it’s hard to predict Cainiao’s fate.

And in this week's earnings announcement, Alibaba shared that it is no longer going to spin off AliCloud into a separate publicly traded company. When asked during the earnings call whether this pan is temporary or permanent, Joe Tsai deftly answered the question by not answering the question. Sounds pretty permanent to me.    

“1+6+n” is now “1+5+(n-1)”.

This latest backtracking on AliCloud’s spinoff sent $BABA stock price plummeting, even with the SEC’s Uptick Rule or Rule 201 triggered, which limits short selling in volatile trading situations. In fact, I feel bad for its communications and investor relations teams – dealing with all this back and forth could not have been easy. 

AliCloud’s spinoff cancellation is the latest, and arguably most consequential, development that’s affecting the company’s prospects. It also happens to be the one business unit I care about and have watched closely over the last three-plus years. So let’s look at whether this latest chapter of reorg hot mess makes sense or not.

Reasonable Reasoning?

The official reason given for canceling the spinoff is: “in light of uncertainties created by recent U.S. export restrictions on advanced computing chips.” Is this reasoning…reasonable?

Yes and no.

Yes, in the sense that when the Commerce Department announced an expanded export control ban on AI chips to China on October 15, the new rules undoubtedly turned the “small yard, high fence” doctrine into “big yard, higher fence”. Life was already hard for Chinese tech companies, big and small, trying to out-compete (or just not fall too far behind) their American counterparts in building and training foundation AI models. After this expanded ban, things got even harder; AliCloud was no exception. 

No, in the sense that this expanded ban was by no means a big surprise and was well anticipated. Thus, claiming an element of surprise and explaining away the spinoff cancellation on the expanded ban alone lacks credibility. After all, when the prospect of an AliCloud spinoff was first made public back in March, the unit had already been dealing with the first version of the chips ban for six months, which took effect in October 2022. It has since been stockpiling H800 chips, the once ok but now banned Nvidia modified GPUs to its H100 series, along with Tencent Cloud, ByteDance, and all its other peers. So it is hard to believe that Alibaba itself didn’t see this coming. 

So what else could explain this sudden reversal? Financial engineering gone awry.  

When asked about this during the earnings call Q&A, Joe Tsai said this:

“About the – our announcement to not proceed with the full spin-off, for us, when we announced the full spin-off, we were looking at a weighted sort of a financial engineering way to show the value of the business…But the circumstances have changed. And right now, rather than focus on financial engineering, we rather focus on figuring out how to grow the cloud business.”

Giving credit where credit is due, Tsai is at least honest and transparent about it. The spinoff was a financial engineering play all along, not a new corporate arrangement aimed to improve and boost the growth prospect of an otherwise promising technology platform. This play fell apart when the cloud unit’s revenue, excluding services provided to other Alibaba units like Taobao and Cainiao, declined this quarter. Based on our analysis, this is only the second time the cloud unit suffered this type of revenue decline, since Alibaba began breaking down revenue per each of its six major units in May 2019.

For the sake of argument, in an imaginary world where AliCloud had somehow completed its spinoff and independently listed on say the NASDAQ or HKEX, releasing an earnings report that shows revenue decline for a major cloud provider would send its share price down the drain. 

Sadly, you can’t financially engineer away declining growth.

A “Value” Stock Future?

Alibaba is stepping into the classic identity crisis of a mature-ish tech company, where it has to straddle a growth narrative and a shareholder value narrative.

These two narratives are quite often in conflict with each other. This conflict was evident in the narrative of its earnings release. On the one hand, executives mentioned multiple times the need to use its financial strengths to make more investments in an AI-driven world to capture this new innovation wave and achieve more sustainable growth. On the other hand, the company approved an annual dividend to shareholders for the first time in its history, worth roughly $2.5 billion USD. This dividend is on top of the company’s share repurchase program, which still has $14.6 billion left to spend.

Dividends usually signal a lack of investment ideas or growth trajectories, thus giving money back to the shareholders is the best use of that cash – the value narrative. If there are promising investment areas or growth trajectories, then aggressively putting that cash to work instead of issuing dividends is the path forward – the growth narrative. Even the most cash-rich tech companies may not pay a dividend, if growth is the focus. Neither Google nor Amazon issues dividends. 

Both narratives are perfectly fine, depending on which direction the company wants to go. But it is nearly impossible to tell both narratives credibly at the same time.

What is not narrative are the numbers on a balance sheet. As of the end of September 30, Alibaba’s cash or cash equivalents stand at $85.6 billion USD or 42% of its market cap. That’s a big pile of dry power. And it is not clear at this moment if Alibaba knows what to do with it. 

As John Maynard Keynes once said, “When the facts change, I change my mind.” I’ve shared in our Q3 2023 performance letter how we’ve been thinking about and assessing Alibaba. Well, facts have undoubtedly changed since I published that letter. In light of its reorg hot mess, we will most definitely be reassessing our view of Alibaba and change our minds accordingly. 



当阿里巴巴在今年三月提出其所谓的 “1+6+n” 企业重组计划时,将这个巨头分解成一个控股公司,加六家相对独立、关系较为松散的业务单位,再加上一组其他资产(即“n”),这可谓是商业史上最大手笔的一番重组计划。当时,我对阿里未来的这一举措给出了一个正面积极的和一个负面愤世嫉俗的解释。由于我天生乐观,最终觉得正面积极的那一面更有可能。

可是到目前为止,这一重组是一团糟,进展并不顺利,证明了负面愤世嫉俗的那一版更有预见性。自从最初的重组官宣以来,当时的CEO张勇从领导控股公司和阿里云,到仅担任阿里云的CEO,再到完全辞职,今后只用公司的钱做些风险投资。蔡崇信成为了集团的主席。吴泳铭成为了集团的CEO。吴还从张勇那里接管了阿里云,但仅是暂时的。盒马,“n”的其中一个成员,本来准备上市,随后计划被搁置。菜鸟虽已提交了在香港上市的有关文件,但尚未定价。鉴于今年全球整体IPO市场面临的困境(见Arm, Instacart, Klaviyo上市后的股价),菜鸟的命运难以预测。


“1+6+n” 变成了“1+5+(n-1)”。

阿里云的分拆被取消使得阿里的股价暴跌,即使在SEC的涨价规则(或"Rule 201")的触发下,也控制不住(规则限制了在大波动交易情况下过度卖空)。我很同情其传媒和投资者关系团队成员门的恼火 —— 处理这些出尔反尔的决策一定不容易。

阿里云取消分拆是最新的,也可以说是最具影响力的发展,直接影响到公司的前景。它也恰好是我密切关注了三年多的一个业务线。那么,让我们来看看这最新一章的 “重组混乱” 的长远意义。




充足的点,在于美国商务部10月15日宣布针对中国控制出口更多种类的AI芯片时,此新规则无疑把所谓的 “小院子,高围栏” 政策方针,变成了 “大院子,更高围栏”。对于中国所有的大中型科技公司来说,日子已经很艰难了,都在试图在构建和训练基础AI大模型方面超越(或至少不要落后太多)它们的美国对手。这次扩大禁令后,情况变得更加严峻;阿里云也不例外。


那还有什么其他原因可以解释这种突然的逆转?“金融工程”(financial engineering)出了问题。









通常,分红利给市场的信号是缺乏投资方向或增长轨迹,所以才把资金返还给股东 —— 即价值故事。如果有很多好的投资方向或增长轨迹,那么积极的把资金投到那些方向,而不是发放红利把钱还给鼓动才是正确的道路 —— 即增长故事。即使是现金最饱满的科技巨头,如果重点是增长,也是可能不发红利的。谷歌和亚马逊都不发放红利。



正如著名经济学家凯恩斯(Keynes)曾经说过的:“当事实改变时,我也会改变我的想法。” 在我们的2023年第三季度业绩信中,我分享了对阿里的思考和评估。自从发表那封信以后,事实无疑发生了变化。鉴于阿里重组的一团糟,我们肯定会重新评估对阿里的看法,并会相应地改变我们的想法。