During the COVID market boom in the latter half of 2020, META announces massive investment into metaverse touted as the next frontier / form in social media platform while spending like there's no tomorrow. Instead of testing it on a microscale with discipline capital injection while assessing market willingness to pay for it, they went in big. Rightfully, they got punished by clear-minded investors/shareholders. Today we know that shareholders were right.
Alibaba foray into cloud industry has been nothing short of spectacular. It was fast & furious with heavy resources & manpower investment & strong execution. In no time, it catches up with global top LLM company eg. OpenAI. Currently, Qwen 3.0 ranks #4 among global open & close - sourced LLM models right behind OpenAI, Gemini 2.5 (Google) & Grok. Latest Q4 earnings release posted an +18% growth charting positive trajectory in terms of growth rate. QWEN has shown early signs of increased enterprise cloud adoption. It's presumptuous to claim AI LLM produces no value (as compared to metaverse) however, a look into recent development in the space tells a murky outlook noted by low barriers to entry. A group of university students of various disciplines led by a venture capital focused on trading algorithms created a low cost LLM that rivals the likes of OpenAI on major benchmarks (Deepseek).
In a previous turnaround Quarterly ER, Joe Tsai, Alibaba's chairman announces heavy investment into AI space of up to USD53b within the next 3 years. Albeit, coming from a PE background with a strong track record, one could tell that they were discipline & level-headed in their investment into AI.
The current Alibaba corporate direction is to stabilise market share in their e-commerce business while fully focusing on developing state of the art world leading AI LLM driven enterprise cloud business with its own ecosystem of enterprise developers centered around it's cutting edge Qwen family of LLM's such as Manus AI agent.
While they fiddle around with a shining new toy, AI, others are eating their lunches. This quarter marks the first time Baba underperformed JD in terms of top line in effect becoming the sick man among the e-commerce platform taking over JD's place.
In recent quarterly ER, JD clocks in overall revenue growth of +15% with ad revenue growing ~ +14% while transactions cost growing ~ +16%. JD's management even touted double digit growth in user base. On the other hand, Baba's TTG group's main monetization model (QZT) via customer service revenue grew around +12%. Overall, TTG revenue grew +8%.
AI cloud business unit is still not profitable while it couldn't even pay for their G&A fees.
Currently AI + Alicloud segment brings in abt 12% of revenue (~USD4.1B). The business unit is profitable at EBITDA level (*not truly profitable since it doesn't include depreciation charges which is main cost in AI Cloud business *just imagine the cost of NVIDIA chips etc) generating about USD 333M (EBITDA) while G&A expenses clock in at USD1.423B for the quarter. It should also account for the majority of the USD2B R&D cost due to aggressive LLM research this quarter.
While Baba invests multi-billion dollars in AI Cloud, it achieved no material improvements so far in its e-commerce services metric as proven by their market lagging growth where AI is supposed to help in user experience eg. Better search recommendation, customer services etc.
One of the reasons is that, all the other players are also doing the same thing, just not with their own $$. They simply subscribe to other AI LLM service providers for peanuts eg. Deepseek thereby preserving their margin in the near term with exception of Bytedance(tiktok).
So far, Alicloud IAAS model borrowed from Amazon hasn't been profitable for Alibaba. Their traditional biggest spender on Alicloud platform, Bytedance/tiktok is now a competitor & edutech industry is no longer what they used to be after Chinese gov politicised them to be bad for living standard & lifestyle. While AI LLM proves to reaccelerate cloud adoption, it has a very low barrier to entry as observed from the myriad of tom ** & harry LLM appearing everywhere.
Baba's latest foray into instant commerce & partnership with Xiaohongshu comes as a desperate move to stem further market share losses from latest April market share data. This shows that management isn't close to the ground & only reacts when data comes in from third party providers. Ideally, we want someone with ears on the ground. This bodes poorly on the management. Alibaba elite management needs to operate out of China instead of New York to stay close to the ground & react on demand to market forces otherwise it'll prove to be fatal.
If Baba continues with their current direction to stabilise Market Share, there's no guarantee they'll drop out of the AI race due to cost constraints even before achieving profitability. This leads to credit rating drops due to poor liquidity metrics. Currently, benefit to risk ratio asymmetrically points steeply downwards for Baba.
It triggers alert when management starts highlighting SBC in their expenses like it doesn't matter. I hope they don't do that anymore because SBC is a business expense at the expense of common shareholders albeit by aligning employees interest to shareholder's.
They should stop using EBITDA for their e-commerce business as it's profitable for the longest time. No one is trying to see if it'll be profitable in the future. 🤦
In the west, Amazon e-commerce business has a very poor margin (*very much JD'ish) but cloud is extremely profitable. In China, people simply don't see enough value to pay for ROM spaces & networking services (something they can easily access while prices drop with time). Plus, China's nascent SaaS industry still isn't ripe for monetization.
If e-commerce isn't in the picture, it's hard to imagine how AI cloud could create enough value to compensate for the slowdown in e-commerce business. Much like imagining Amazon's e-commerce arm growing enough profits to cover shortfall from cloud underperformance.
in that case, while cloud AI rev trajectory is in the right direction, growth at +18% is insufficient. They need to be at least growing at >100% to be considered fair at this stage given the relatively small size of Cloud Intelligence arm & the lagging performance in the e-commerce arm.
Alas, consider it a glimpse of hope that Xiaohongshu (little Redbook) has integrated into Ali Cloud. Fortunately, management already made some arrangements with Little Rednote(Xiaohongshu) starting with a small collaboration via links to taobao. It's small likely because of high premium of Xiaohongshu & political-business environment in China. If results of small collaboration were great, I believe it'll go through to a bigger form of collaboration or even a major acquisition if political environment allows.
Baba's collaboration with Xiaohongshu along with instant fulfillment of e-commerce is the only redemption. Baba's management should quickly & urgently engineer a complete integration of Taobao into Xiaohongshu geared towards taking advantage of impulse buying market trend before it's too late (April GMV data shows Bytedance's Douyin clocking north +40% gmv growth almost double of PDD). App experience should be designed to provide all tools & intuitive ease for purchasing whatever users find interesting on Xiaohongshu/Little Rednote making use of machine vision & AI LLM search tailoring based on graphics rather than keywords as currently in use. In fact, this will allow further innovation in monetization based on graphics targeted search rather than keywords.
That's the only way Alibaba could increase the likelihood of survival till its AI Cloud efforts bear fruits for nobody can tell how much investment, how long it'd take to win the AI race & how'd it look like.
Lastly, business rise & fall with the tides of times. Business triumphs & destruction depends on its management & executive's ability to take advantage & adapt to market forces & trends. No one can tell if it'd end up the way META adventure into metaverse.
So on behalf of Baba's public common shareholder, I'd like to end this article with a shout out to Baba's management & executives,
"Wake up & act quickly"
Comments