It has been more than 440 days since Jack Ma’s infamous speech at the 2020 Bund Finance Summit. (We have the full speech translated in English, Chinese, and Japanese. It’s worth a read.)
Since the speech, Alibaba’s stock tumbled by more than 50% over the course of 2021. Many other Chinese tech companies’ stocks fell with a similar velocity and magnitude. The entire technology and investment ecosystem in China got tossed into a whirlwind of regulatory crackdowns and policy changes, much of which is still ongoing.
I’ve been hearing a common sentiment from many investors: “China is uninvestable.” Yet, one (dare I say) pretty good investor begs to differ – Charlie Munger. The investment portfolio of Daily Journal, of which Munger is the chairman, began buying Alibaba shares during H2 of 2021. During Q4 of 2021, its Alibaba position effectively doubled. $BABA is now the Daily Journal’s third-largest holding after Bank of America and Wells Fargo, at a reported price of $118.79.
Anyone who follows the investment career of the Buffett-Munger duo knows they don’t trade stocks for short-term gains; they are long-term, multi-year holders. Thus, whatever the now 98-year-old Munger saw in Alibaba, he has determined that this investment will be worthwhile well beyond his 100th birthday.
What did “Poor Charlie” see in Alibaba, or more likely what did his China-focused protege and confidante, Li Lu, help him see, we may never know. What we can deduce from this investment is that China is far from being “uninvestable”.
I share a similar sentiment, which I articulated in a previous post “Removing China’s ‘Startup Debt’”. 2021 has been a painful but necessary year to clean up 40-plus-years of “startup debt” that the Chinese economy has accumulated since “opening and reform”. Alibaba, being the country’s largest tech company, probably has the most “debt” to clean up.
Whether you are a tech company or a country, “removing startup debt” generally creates a more sustainable foundation to grow for the long term. In the short term, you have to slow down your pace of shipping new features and improving your products, but you can never completely stop.
Despite a tumultuous 2021, Alibaba has not stopped shipping. Most of the analysis on Alibaba is focused on its e-commerce, food delivery, and other consumer-facing products – all of which are low-margin businesses that are slowing down. Alibaba’s long-term moat, growth, and profitability lie in its cloud platform, B2B enterprise businesses, and more “hardcore” technical innovations. That’s where I suspect Munger found Alibaba appealing, otherwise he would have invested in the company many years ago, not 2021 of all years. That’s also where I have seen the most ''ships'' and improvements. Let’s dig into what they are.
AliCloud: A Profitable Cloud
Alibaba has built its cloud platform, AliCloud, into a profitable business. The division has been profitable (on adjusted-EBITA basis) for the last four quarters.
Building a profitable cloud is a rare feat. It takes an incredible amount of forward planning, discipline, operational efficiency, and hedgehog-like execution. It’s a level few in tech has achieved. Actually not few, just one: AWS.
Amazon’s AWS has been profitable since 2015. But Google’s GCP is still bleeding money ($644 million loss in its most recent quarter). Microsoft does not share Azure’s profitability status, so no one really knows. Thus, AliCloud is the only other profitable cloud of comparable size. (To be clear, there are other smaller profitable clouds like Digital Ocean, but their sizes and market shares pale in comparison to the Big Tech’s clouds.)
Granted, AliCloud’s share of Alibaba’s total business is still small – no more than 10%. The business unit also had many things working against it in 2021. It lost a huge customer (rumored to be ByteDance) possibly due to geopolitical tension. It incorporated DingTalk into the unit in April 2021, which is Alibaba’s workplace SaaS product that is still young and mostly given out for free to drive adoption, thus hurting profitability.
What will be worth watching when Alibaba publishes its next earnings results is if AliCloud was able to maintain profitability for the entirety of a rough 2021. If so, then a consistently profitable AliCloud, especially during lean times, can serve as the long-term foundation for global expansion of Alibaba’s enterprise business and more “hardcore” technical innovation – both of which take a long time to pay off and both of which are already underway.
APAC-Wide B2B Expansion
With the antitrust regulatory landscape in China still in-flux, it's hard to predict how much Big Tech’s like Alibaba are allowed to grow domestically. Thus, overseas expansion beyond China has become a higher priority. Expansion also aligns better with the government’s overall objective to keep growing China's overall influence.
While the US market is effectively off-limit for AliCloud, other emerging markets are wide open spaces. In a previous post, “Where Can the Chinese Internet Go?”, I've identified Southeast Asia, Latin America, and the Middle East as three top regions where AliCloud (and other Chinese cloud platforms) can grow. But cloud expansion is an expensive endeavor, so having an already profitable business that can sustainably make investments is crucial.
AliCloud seems to be doing exactly that, particularly in the APAC region. In the last two years, it has announced new data centers in South Korea, Thailand, the Philippines, and Indonesia. It is building small businesses, startups, and developer ecosystems around AliCloud, targeting the Philippines, Thailand, and South Korea. It won an official partnership to provide cloud computing services for the Malaysian government (along with Telekom Malaysia, Amazon, Google, and Microsoft, so the deal is not exclusive), while announcing the opening of its first-ever non-China-based innovation center to be located in Kuala Lumpur. (These “innovation centers” are often used as “sweeteners” to lock in a government contract in emerging markets.)
AliCloud has not stopped making moves, though mostly under the radar, devoid of major media attention. (Western mainstream media prefer to write non-stop about "where is Jack Ma".)
Targeting APAC as AliCloud’s next B2B growth vector makes sense due to three factors:
1. The region’s proximity to China and its sphere of influence;
2. Alibaba’s existing e-commerce traction in most APAC markets;
3. A macro trend that APAC has the largest and fastest-growing developer community – a prerequisite to cloud business growth.
Factor #3 is a reality that most American tech companies have yet to grasp or invest in heavily, leaving a wide-open space for AliCloud to execute against.
“Hardcore” Tech and Open Source
In addition to market expansion overseas, a lot of “hardcore” technology innovation is happening within AliCloud. I’m calling these innovations “hardcore”, because they are both difficult to achieve technically and mostly related to hardware. Here are a few made public in 2021 that are worth highlighting, spanning all layers of the stack from chips and processor, to operating system and database:
Yitian processor: in its September 2021 quarterly earnings report, Alibaba announced its home-grown Yitian 5nm processor, which it plans to deploy into its own proprietary servers inside AliCloud’s many and growing data centers. Building your own cloud server processor is no joking matter. AliCloud engineers appear up to the task. Yitian is supposedly the fastest Arm processor on the market. While replacing processors, in this case probably Intel’s, with your own won’t yield much business impact overnight, deploying Yitian is key to increasing AliCloud’s profit margin over time, not to mention reducing dependency on an American supplier that may get caught up in future sanctions.
Xuantie: this is Alibaba’s own semiconductor design, which leverages RISC-V, the dominant open-source instruction-set architecture. It is still a young project. What’s telling is Alibaba’s willingness to open source it so early in its development, showing that the company recognizes the value of open source in its long-term strategy and the positioning it could gain among proprietary chip architectures, like Intel’s x86 and Arm.
AnolisOS: this is a home-grown operating system that has apparently been in development for 10 years and serves as the OS for the system that makes Singles Day possible. The newsworthy thing is AliCloud decided to open source AnolisOS last year and made it compatible with CentOS, a popular Linux operating system distribution. Linux still dominates the cloud computing server market – even Azure runs more Linux servers than Windows servers. I don’t see Anolis taking any meaningful market share away from Linux. What this news does show is the AliCloud engineering team’s confidence in their own product and, again, their willingness to embrace open source, create an ecosystem around Anolis, and eventually make AliCloud’s stack more home-grown.
PolarDB-X: this is AliCloud’s main distributed database offering and the combined product of previously two separate database units, called PolarDB and X-DB (thus the PolarDB-X name). Similar to Anolis, this database has been in development for a few years now, but the newsworthy element last year was, again, the announcement to open source PolarDB-X. A highly performant and reliable distributed database is extremely hard to build. Yet, every cloud needs its own flagship distributed database offering to be successful. For AWS, it is Aurora. For Azure, it is Cosmos. For GCP, it is Spanner. For AliCloud, it is PolarDB-X. Interestingly, only PolarDB-X is open-sourced.
One pattern you’ve probably observed at this point is that AliCloud’s “hardcore” technology development is increasingly intertwined with its willingness to open source them, as a way to gain market share versus competitors and pave the way for long-term profitability. As long-time readers of Interconnected know, we’ve written many posts about the intricacies of open source, both as a technology development model and an important element of an overall business strategy. In short, it takes a long time for open source to pay off, but once it starts paying, it pays in spades. Given its current profitability, AliCloud is well-positioned to reap the fruits of its own innovation via open source in the long run.
Does Charlie Munger understand or care about the specifics of any of these “hardcore” tech or the intricacies of how open-sourcing them can contribute to Alibaba’s overall growth? Probably not. But what Alibaba has been doing during a rough 2021 was continuing to invest, expand, ship, and build its moat on the back of a profitable AliCloud – that much Munger definitely understands. And as Alibaba and China as a whole continue to remove their “startup debt”, he, along with Daily Journal shareholders, stands to benefit from it all for years to come.
Disclosure: I do not own any BABA shares. I did during most of 2020 and 2021, but sold them all recently to fund other projects. Maybe Munger bought them. That’s why I’m not as good of an investor as he is; very few people are.
我从许多投资人那里听到一个共同的观点就是："中国不可投。" 然而，一位（我敢说）成绩相当优秀的投资人却不敢苟同，他就是查理·芒格。芒格担任主席的Daily Journal 在2021年下半年开始狂买阿里股票。在2021年第四季度，其阿里的持股总数翻了倍。$BABA现在是 Daily Journal 继美国银行和富国银行之后的第三大持股，报告价格为118.79美元。
我也有类似的观点，曾在之前的一篇文章《清除中国的"创业债"》中阐述了这一点。2021年是痛苦但必要的一年，以清理中国经济自 "改革开放" 以来积累的40多年的 "创业债"。阿里作为中国最大的科技公司，需要清理的"债" 也可能最多。
尽管2021年很动荡，阿里并没有停止创新改进。对阿里的分析大多集中在其电商、外卖配送和其他to C消费者的产品上 —— 这些都是薄利的业务，增长正在放缓。阿里的长期护城河、增长点和盈利能力在于其云平台、B2B企业业务和更多的 "硬核" 技术革新。这就是我怀疑芒格发现阿里吸引人的地方，否则他多年前就会投它了，而不是等到2021年。这也是我看到最多创新和改进的业务方面，让我们来看看创新改进的都是什么。
阿里已经将其云平台 —— 阿里云 —— 打造成为了一个盈利的业务，该部门在过去四个季度里一直在盈利（按调整后的EBITA计算）。
诚然，阿里云在阿里总业务中的份额仍然很小 —— 还没超过10%；该业务部门在2021年也不走运，失去了个大客户（据传是字节），可能是因为地缘政治紧张的关系。阿里云在2021年4月还将钉钉纳入该部门，一个还很年轻的产品，大部分是免费送客户来推动使用，因此是亏损的。
当阿里公布其下一份财报时，值得关注的是阿里云是否能够在艰难的2021年中一整年保持盈利。如果是的话，那能够持续盈利的阿里云，可以作为阿里企业业务全球扩张和更多 "硬核" 技术创新的坚实后盾和长期基础 —— 这两者的回报都需要很长时间，而它们也都已经在进行中。
3. 亚太地区拥有最大和增长最快的开发者社区这一宏观趋势 —— 这是云计算业务增长的先决条件。
除了海外市场的拓展，阿里云内部也在进行大量的 "硬核" 技术革新。我称这些创新为 "硬核"，是因为它们既在技术落实上难度很高，又大多与硬件有关。以下是2021年公布的几个值得关注的创新，横跨从芯片和处理器，到操作系统和数据库的整个技术栈。
AnolisOS：这是一套自产的操作系统，显然已经开发了10年，作为运作“双十一”的系统的OS。有新闻价值的点是阿里云去年决定将AnolisOS开源，并使其与Linux操作系统发行版CentOS兼容。Linux仍是云计算服务器OS份额的最大一块 —— 就连在Azure上运行的Linux都比Windows要多。我不认为Anolis会从Linux手中夺走任何有意义的市场份额，但AnolisOS开源这一消息显示的是阿里云工程团队对自己产品的信心，以及他们的意愿去拥抱开源，围绕Anolis打造生态，并最终使阿里云的技术堆栈能使用更多自制的产品。
您可能已经察觉到了一个规律，那就是阿里云的 "硬核" 技术开发与它拥抱开源的意愿日益交织在一起，以此从竞争对手那里抢到市场份额，并为长期盈利铺平道路。《互联》的长期读者都知道，我们写了很多分析开源方方面面的文章，它既是一种技术开发模式，也是一套商业战略的重要元素。简而言之，开源需要很长的时间才能得到回报，但一旦回报开始，成绩是显赫的。鉴于其目前的盈利能力，阿里云可以通过对开源的投资而从自己的创新中长期受益。
查理·芒格了解或关心这些 "硬核" 技术的具体细节，或开源会如何促进阿里的整体增长的复杂关系吗？应该不的。但是阿里在艰难的2021年中仍在继续投资、扩张、出新产品，并在盈利的阿里云内继续挖自己的护城河，这一点芒格是很清楚的。随着阿里和整个中国继续清除其 "创业债"，他和Daily Journal的股东们将是未来多年的受益者。