Instant Retail and AI: Alibaba’s "Fragmented Narrative"

The investment allocation between AI infrastructure and instant retail has become the key to "balancing" Alibaba's valuation.

Currently, there are generally two mainstream opinions in the market regarding Alibaba's transformation towards AI.

One benchmarks it against Google. Morgan Stanley pointed out in a report in March this year that "Alibaba is the Chinese technology giant closest to Alphabet (Google)'s vertical integration capabilities, spanning three layers: chips, cloud, and basic models." Goldman Sachs held the same view in its report in November last year, believing that "Alibaba is adopting a 'full-stack' approach similar to Google, betting on self-developed chips, cloud, and basic models."

One benchmarks it against Amazon. JPMorgan Chase pointed out in its report in April this year that Alibaba is first and foremost an e-commerce + cloud company, exactly the same as Amazon; AI is a long-term option, not a short-term driver, "We use Amazon's peer multiples as the valuation anchor." HSBC held a similar view in its report in March.

If calculated based on Google's 25x P/E ratio, Alibaba's stock price has an upside potential of 40 - 50%. The imagination space of the Amazon narrative is also not bad, but its P/E ratio is close to 40 times, and Alibaba's e-commerce growth rate is lower, so it's difficult to reach the same multiple. More importantly, if AI is regarded as a "long-term option," the valuation will be infinitely suppressed.

In comparison, the "Google narrative" is obviously more cost-effective for Alibaba's valuation potential.

To tell the "Google story" well, Alibaba has taken a series of intensive actions, including significantly increasing Capex, and aiming to complete the full-stack self-developed closed-loop of "chips - cloud - models," promoting the formation of the consensus that "only Google and Alibaba in the world have achieved full-link self-development." All these efforts have been warmly welcomed by the market.

However, beneath the "Google narrative," the role of the e-commerce business is changing.

Recently, Alibaba announced that Wu Zeming has been newly promoted as one of the five partnership committee members, driving the stock price to soar. On the one hand, the market recognizes that its AI leadership has gradually taken shape; on the other hand, the AI strategic faction and the retail business faction have formed distinct camps, gradually moving from an attempt at synergy to a parallel narrative.

Hidden Line of Power and Responsibility

In the past year, the Qianwen large model has had a total of six major iterations, with five intensive iterations in the first half of 2026. In the first five months, there was an "update every month," and each iteration drove the stock price up. Roughly calculated, the Hong Kong stock rose by 2% - 4%, and the US stock rose by 1.5% - 3.5%. The impact of the implementation of agents such as the release of Wukong and Dianxiaomi on Alibaba's stock price increase was relatively mild, mostly between 1% - 2%, significantly reflecting the market's preference.

However, compared with technological breakthroughs, the hidden line of organizational structure adjustment and personnel changes has a more significant impact on the valuation in the "Google narrative."

On March 4, 2026, Lin Junyang announced his departure. On that day, Alibaba's Hong Kong stock fell by nearly 5.9%, and the US stock fell by 5.5%. On March 16, ATH was announced to be established, and the Hong Kong stock rose by 7.8% in response, and the US stock rose by 7.5%.

Among them, "Whether the technology faction can gain the upper hand" has become one of the focuses.

On June 2, it was reported that Wu Zeming (Fan Yu), the CTO of Alibaba Group, officially joined the Alibaba Partnership Committee in May, ranking among the five members of the committee together with Ma Yun, Cai Chongxin, Wu Yongming, and Jiang Fan. At the same time, Yan Xiaolei, the CEO of Hema, changed from reporting to Wu Zeming to reporting directly to Jiang Fan.

After AI was completely upgraded from a "business line" to a "top-level strategy," on that day, Alibaba's Hong Kong stock closed up 6.60%, and the US stock closed up 4.32%.

This trend was practiced as early as April. At that time, Wu Yongming said in an internal letter that Wu Zeming (Fan Yu) would focus on the CTO work of Alibaba Group, be responsible for the construction of the group's business technology platform and AI inference platform, and serve as the convener of the technology committee. His position as the CEO of Taobao Flash Sale was taken over by Lei Yanqun.

The market interpreted this adjustment as "the technology faction regaining the leadership of AI." On that day, Alibaba's Hong Kong stock closed up 6.75%.

These two purifications of power gradually separated Wu Zeming from the business circle under Jiang Fan's control and officially incorporated him into the core of the technological power structure built by Wu Yongming. Reflected in the valuation, the positive impact of this personnel adjustment on Alibaba's stock price can even be comparable to the establishment of the ATH division, and is significantly higher than structural adjustments such as the integration of the Token Foundry division.

This means that the market has extremely high expectations for the promotion of the leadership of the technology faction at the top within Alibaba, which is even directly related to whether Alibaba can continue to tell a good AI narrative similar to Google's.

"Fragmented" Narrative

Unfortunately, in this narrative, the situation of the e-commerce business is a bit awkward.

Although in Alibaba's statement, the two are "in the same logic" - Alibaba Chairman Cai Chongxin once said in the Q3 financial report of 2026 that "burning money on instant retail is not for short-term profit, but for building the infrastructure, data, scenarios, and user mindset in the AI era"; Alibaba CEO Wu Yongming continued to emphasize in the Q4 financial report that "we now have a closed-loop of data - model - scenario - more data, and instant retail is an irreplaceable engine in this closed-loop."

Among them, high-value "data" is indeed a part that Alibaba is extremely eager for at present - although it has a huge business ecosystem, in terms of key indicators such as data dimension, data frequency, and data liquidity, Alibaba still has a lot to learn compared with Tencent and ByteDance in terms of data richness. This is also the main reason why Qianwen was pushed to the forefront and became the pre-container for data absorption.

However, the market doesn't seem to buy it. In May 2025, after Alibaba announced an investment of 50 billion yuan in subsidies to expand Taobao Flash Sale, the stock price fell by a cumulative 23% by July; after confirming another investment of 50 billion yuan in subsidies in the next 12 months in July, the stock price hit a new low; in March and May 2026, the management successively announced continued investment in instant retail, causing the stock price to plummet twice.

"Alibaba is facing a continuous tension between technological idealism (AI/cloud) and business reality (instant retail subsidy war). This internal contradiction, namely the conflict between high AI investment and aggressive retail spending, defines the risk-return characteristics of the stock and is the biggest internal contradiction for Alibaba at present," pointed out Jefferies in its report "Alibaba: MaaS to Drive Next 5 Years" released in April 2026.

Morgan Stanley also pointed out in its report "Best AI Enabler in China" in May 2026 that "investors essentially price AI as a growth option and instant retail as a profit put option. The tug-of-war between long-term AI ambitions and short-term retail losses remains the core factor suppressing market sentiment."

Alibaba has also tried to bridge the gap between the "Google narrative" and "substantial subsidies."

Initially, it tried to clarify the logic of "instant retail is valuable, very valuable, and will become more and more valuable for AI." As Jiang Fan, the CEO of Alibaba's e-commerce business group, said in the Q1 financial report of 2026, "In the next three years, flash sales and instant retail are expected to bring an additional 1 trillion yuan in transactions. These high-frequency, highly real-time, and high-scenario-density transactions will become the most important training data source and application scenario for our AI models."

At the same time, the main line cannot be abandoned. At the Yunqi Conference in 2025, Wu Yongming said that Alibaba's strategic focus is "user first, AI-driven," and it aims to "transform from a traditional e-commerce platform into an open technology platform enterprise serving AI innovation in the whole society."

If we elaborate on the positioning of the e-commerce business, Wu Yongming's statement at the Yunqi Conference is that Alibaba will be driven by two wheels in the future. One is the consumer platform, responsible for cash flow and the basic market; the other is the AI + cloud technology platform, responsible for future growth and value creation.

To put it simply, "e-commerce is responsible for cash flow, and AI is responsible for valuation and future growth."

The "Turnaround" of the E-commerce Business

According to the annual report for the fiscal year 2026 released by Alibaba in May, the adjusted EBITA of Taotian (Chinese e-commerce) was 107.5 billion yuan, a year-on-year decrease of 44%; the group's free cash flow was -46.6 billion yuan, a year-on-year decrease of about 120.48 billion yuan, a decline of about 163.1%.

If we add the annual AI infrastructure investment of about 126 billion yuan (corresponding to an investment of 380 billion yuan over three years) and the estimated instant retail loss of about 70 billion yuan by multiple securities firms, the total strategic expenditure for the entire fiscal year is about 200 billion yuan, almost all covered by Taotian's profits - during this period, the adjusted EBITA of Alibaba Cloud was only 14.27 billion yuan, and both AIDC and other businesses were in the red.

Moreover, the group's R & D expenses in the fiscal year 2026 increased by about 16% year-on-year, mainly invested in AI computing power and cloud infrastructure, while the growth rate of the core customer management revenue (CMR) of the e-commerce business was only 1%.

To become like Google, Alibaba still needs to continue to build a system-level entrance, achieve global coverage, and even lead in basic research globally in the long term, forming a global technology infrastructure and a "dual-ecosystem closed-loop of information + business." The cloud business cannot shoulder this responsibility, and other businesses are even less reliable. It still has to rely on the e-commerce business to continue to "transfuse blood."

One can imagine the mood of Jiang Fan, the CEO of the e-commerce business group.

Morgan Stanley also added fuel to the fire in its report "China’s AI Path: Owning the Full AI Stack" released in September 2025: "Alibaba's cloud/AI business is the core valuation engine, while instant retail (QC) remains the main cash consumer in the short term. Striking a balance between large-scale AI capital expenditure and aggressive instant retail subsidies will be a key test of the management's fund allocation discipline."

Facing the choice of whether instant retail is an "AI testing ground" or a "business battlefield," Jiang Fan said in March 2025 that "first gain enough market share, and then talk about profit and value realization," and in November, he said that "optimizing UE will be the core goal of instant retail."

In April this year, after Lei Yanqun took office, he issued an all-staff letter saying that he would comprehensively reduce subsidies with low conversion rates and low gross margins; significantly shrink large-scale full discounts/free orders; the annual core goal is to halve the loss (from about 70 billion yuan to 35 billion yuan). This is the first time that Alibaba has quantified the loss reduction target of instant retail to a specific amount, marking that its "pursuit of profit" has transformed from a strategic goal into a hard assessment.

At the earnings call for the fiscal year 2026 on May 13, Jiang Fan for the first time clearly stated the specific time for UE to turn positive, "We are confident that we will achieve positive UE by the end of the fiscal year 2027. We will achieve overall profitability in the fiscal year 2029."

Valuation Direction

With Wu Zeming joining the partnership committee, the narratives of AI infrastructure and instant retail seem to have become more parallel.

Let's take a look at the promotion resume of this "veteran technology faction." Since 2022, Wu Zeming has been appointed by Zhang Yong as the group's CTO, and at the same time, he also serves as the CTO of local life, the CTO of Taotian, and the chairman of Ele.me; from 2024 - 2025, he concurrently served as the CEO of Taobao Flash Sale and personally led the fight in the food delivery war.

At that time, technology and business were still closely integrated in Alibaba. In the previous food delivery war, Wu Zeming was also responsible for driving business growth with technology (AI).

After this adjustment completely separated him from the business line, the formations of the AI strategic faction and the retail business faction have become clearer. On one side are Wu Yongming, Zhou Jingren, and Wu Zeming; on the other side are Jiang Fan, Lei Yanqun, and the old iron army.

Since the logic of "instant retail is valuable for AI" is not recognized, the strategic priority of the e-commerce business has sharply declined. However, building an AI narrative requires profit support, so Jiang Fan and Lei Yanqun have taken more initiatives.

On June 2, Wang Xing, the CEO of Meituan, repeatedly emphasized in the Q1 earnings call that the competition situation of food delivery subsidies has begun to become more rational. On June 3, Alibaba indirectly stated externally that "Meituan is not invincible, and the food delivery war has just begun."

It is logical to shift from cooperation to pursuing profit, which is also equivalent to announcing that the attempt at synergy between e-commerce and AI has been temporarily terminated. Therefore, in the short term, the market may not see the results of Alibaba's efforts to bridge the gap between the two fragmented narratives of AI and business. This will further test the elasticity of Alibaba's valuation.

The question still remains to be answered. In the current prosperous AI trading environment, in what direction should Alibaba be valued?

This article is from the WeChat official account “Mingliang Company” (ID: suchbright), author: Soda. It is published by 36Kr with authorization.