Alibaba’s vision for the agentic era comes into focus as it targets $100B in AI and cloud revenue over 5 years 

While it started out as a B-to-B marketplace, Alibaba’s sprawling business now encompasses everything from consumer-facing apps like Taobao and Tmall to cloud computing to enterprise AI agents. 

And it’s the latter two segments that Alibaba is counting on to drive a lot of growth in the near future. Specifically, Alibaba is projecting that revenue from its AI and cloud computing divisions will hit $100 billion over the next five years. 

But to get there, it’s spending a lot of money and moving quickly to push heavy adoption of its new consumer-facing and enterprise applications. Last year, the company pledged to invest at least $53 billion in AI infrastructure over the next three years.

Earlier this week, it was announced that Alibaba was reorganizing its AI operations under a new business unit called Alibaba Token Hub. The news came just days after it was reported that one of Alibaba’s star AI researchers, Junyang Lin, was stepping down from one of its marquee projects. 

As such, there were a lot of questions heading into Thursday’s earnings call about what this all meant for Alibaba’s AI strategy. During the call, Alibaba Group CEO Eddie Wu spoke about how some of these moves — like the establishment of the Alibaba Token Hub — were meant to help position Alibaba for success amid the current “agentic AI era.” 

During the earnings call, Wu spoke about how — during earlier eras of AI development — large language models, or LLMs, were largely trained by static sets of data. That has changed in the agentic AI era, where companies like Alibaba and Amazon are racing to develop agentic AI bots that can act on their own more autonomously and complete commerce transactions with little to no input. It requires LLMs to be trained on a steady stream of current consumer data, and that, according to Wu calls for closer integrations between applications and LLMs. 

“I think what’s most different and most important about the agentic AI era is the need to achieve this tight integration between the application and model. That’s the critical priority.” Wu said. 

At the heart of Alibaba’s AI strategy is its series of large language models, called Qwen. The first open-source Qwen models were released in 2023. According to some accounts, Qwen is the world’s most widely used open-source AI system. 

In November, Alibaba also released a consumer-facing Qwen app, built off of the most advanced version of its Qwen LLMs. Alibaba has moved quickly to update this app and add more commerce capabilities over just a few short months. In January, it added the capability for people to order food and make travel plans, among other tasks. 

In February, Alibaba said it planned to spend about $431 million during the Lunar New Year holiday to attract users to Qwen, according to Reuters. 

As part of the push to encourage adoption, Alibaba offered consumer incentives, including a bubble tea promotion that spurred users to claim more than 10 million free drinks, driving a surge in downloads and overwhelming some shops across China, South China Morning Post reported.

The subsidies appear to have paid off. Qwen handled nearly 200 million orders during the Lunar New Year period, Alibaba said on its official WeChat account.

Unlike U.S. companies such as OpenAI and Google, Alibaba has structural advantages because it owns a large network of consumer services that its AI shopping app Qwen can plug into, like Taobao and Tmall. Alibaba also operates the Alipay payments network through affiliate Ant Group, logistics provider Cainiao, mapping service Amap and travel platform Figgy. Put together, that means Qwen could theoretically help consumers find a product, buy it, pay for it and ship it inside a single chatbot interface.

Turning AI systems into reliable shopping agents remains technically difficult, in part because these assistants need large amounts of structured data not just to recommend products but also to complete purchases, as Wu’s remarks during the earnings call hinted at. 

E-commerce catalogs, prices and inventory levels are constantly changing, and if that data is incomplete or inconsistent, an AI system could recommend the wrong item, show an outdated price or attempt to buy something that is out of stock. The Information reported in January that OpenAI has been slow to roll out checkout features inside ChatGPT, partly because retailers’ product data is often inconsistent or poorly structured.

According to Juozas Kaziukėnas, an independent e-commerce analyst, Alibaba’s advantages lie in the fact that “Qwen has companies in the universe of Alibaba that it can integrate with without needing external partners, and it has also been willing to aggressively launch, as opposed to waiting for a slow rollout. These two differences alone make it a very different proposition from what Google or OpenAI have been doing.”

“We don’t have a comparable company in the U.S. that would have this full coverage of everything,” Kaziukėnas added. “Perhaps Amazon is the closest, because it does have grocery and normal commerce.”

Enterprise applications will also play a big role in Alibaba’s AI push. Earlier this week, the company unveiled a new AI enterprise platform called Wukong. 

While Alibaba has the advantage of being a well-funded first mover, it is not the only major player pushing to get more people hooked on AI assistants.

In recent weeks, OpenClaw, a personal digital assistant, has surged in popularity among Chinese consumers and businesses. OpenClaw was built by an Austrian developer, but Chinese tech giants Baidu and Tencent have held sessions to help people get OpenClaw set up on their computers. It ties into the Chinese government’s push as a whole to integrate AI into the vast majority of its economy. 

In a March 11 note, Morningstar senior equity analyst Chelsey Tam, who covers Alibaba, said that in light of the departure of Lin, one of the most visible technical leaders working on Qwen, “uncertainty surrounding Alibaba’s AI leadership in China has increased.” But, she added, “conversely, if Qwen will be able to effectively enhance Alibaba’s products and services across the ecosystem through AI, it could strengthen the company’s long-term market position and financial performance.”