PDD AI Large Model Undercover Battle

Author | Huang Yu

Editor | Zhou Zhiyu

Compared to the vigorous layout of AI large models by e-commerce giants like Alibaba and JD.com, the dark horse known as PDD has gone silent.

In the past two years, PDD has not made significant moves in AI large models, and even as domestic and foreign companies are eagerly integrating Deepseek, PDD remains "as steady as a mountain," showing no intention of joining the fray.

However, recent reports indicate that PDD has internally formed a team for e-commerce recommendation large models, led by a core member from Baidu's Phoenix Nest, with an annual salary "several times that of the Baidu period."

As of the time of publication, PDD has not responded to Wall Street Insights. However, Wall Street Insights learned from a headhunter that PDD is indeed still hiring large model algorithm engineers, but places more emphasis on the ability to apply the technology in conjunction with business.

In this wave of AI, referred to as the "Fourth Industrial Revolution," PDD cannot remain inactive.

As a latecomer in the e-commerce industry, PDD has not only carved out a niche in the e-commerce market dominated by the dual giants Alibaba and JD.com, achieving a miraculous comeback, but also surpassed Alibaba in market value for the first time in November 2023. However, at the end of August last year, PDD took a self-inflicted hit, with its stock price dropping nearly 30%, falling far behind Alibaba again.

PDD has moved past the rapid development phase, while Alibaba and JD.com are gradually regaining their footing. In this round of global capital re-evaluation of Chinese tech assets triggered by Deepseek, PDD's valuation has rebounded, but its growth has been relatively modest due to the lack of direct momentum from AI product implementation.

The "AI e-commerce" era that Jack Ma spoke of has just begun, and the competitive landscape of the e-commerce market will be rewritten once again. In the next decade, if the "down-to-earth" PDD wants to maintain its position and not be left behind by the times, it must seize the revolutionary opportunities brought by AI beyond low prices and overseas expansion.

Entry

The last time PDD was reported to be related to large models was in November 2023.

At that time, multiple job postings indicated that PDD was offering high salaries to recruit NLP (Natural Language Processing) large model algorithm engineers to establish an AI large model team and explore applications of AI large models in customer service, dialogue, and other scenarios.

However, even after more than a year, PDD has not presented any mature AI products or features to the public.

PDD has not given up on continuous investment in large model applications, including heavily poaching talent from Baidu last year.

It is reported that PDD currently has several applications: price comparison systems, recommendations, advertising, search, customer service, etc., with each small group's large model team competing against each other.

From the current information, PDD does not prioritize the research and development of underlying large model technologies but focuses on applying AI large models to core e-commerce applications.

The development of PDD's AI large model applications is very low-key, with weak perception among most internal employees as well as externally Multiple PDD employees told Wall Street News that PDD has not actively implemented AI large model applications internally, with the only noticeable aspect being AI intelligent customer service, which may not even qualify as a large model level.

Wall Street News learned from insiders at PDD that while major internet companies have been discussing AI large model empowerment at their earnings calls, PDD has not proactively mentioned any related strategies in the past two years.

In the wave of AI large models over the past two years, PDD can be considered the only major internet company absent from the scene.

As the e-commerce giant, Alibaba's emphasis on AI large models is undeniable, having officially announced its "User First, AI Driven" strategy in September 2023, marking a shift in Alibaba's core strategy in the AI era.

Recently, Alibaba Group CEO Eddie Wu also dropped another bombshell, stating that the company will invest over 380 billion yuan in the next three years to build cloud and AI hardware infrastructure, surpassing the total investment of the past decade. Alibaba is making significant investments to stride into the AI era.

Additionally, several of PDD's competitors in the e-commerce industry—such as JD.com and Douyin—also place great importance on AI large models. For instance, JD.com not only integrates AI large models across multiple business lines but also has its own self-developed "Yanxi Large Model."

Conservatism

The reason for PDD's restrained approach to AI can be traced back to its company culture and founder Huang Zheng's judgment on AI.

In a letter to shareholders in 2018, the year PDD went public, Huang Zheng mentioned that PDD's core value is "conservatism," which means sticking to its own duties.

Against this backdrop, since its establishment, PDD's business scope has always remained focused on e-commerce, consistently dedicating itself to hard work without any inclination to show off.

During the years when PDD concentrated on its e-commerce business, accumulating users, upgrading its supply chain, and enhancing its reputation, Alibaba and others had already shifted their focus beyond just e-commerce, making significant investments in logistics, local life services, new retail, finance, cloud computing, and more.

Therefore, an important reason that cannot be ignored is that Alibaba, JD.com, Tencent, and others view cloud services as a crucial strategic growth point for the future. Investing in large models and generative AI can enhance the profitability of cloud services and open up new business growth opportunities, while PDD, which focuses solely on e-commerce, has not yet expressed such demands.

An institutional analyst told Wall Street News that the current domestic e-commerce competition is fierce, and the market's attention on PDD lies in its overseas business expansion progress.

Minsheng Securities analyst Yi Yongjian also stated that TEMU is the driving force behind PDD's current growth, and the construction of fulfillment capabilities and solutions has become an important influencing factor in Temu's overseas competition.

Moreover, it is well known that the research and development of AI large models is a "money-burning" investment that could divert the company's resources and attention. For PDD, which emphasizes human efficiency, it is clear that it will not easily invest in the research and development of AI large models.

Huang Zheng once held a meeting with all employees after the IPO to look ahead and encourage staff. Regarding this internal meeting, a former PDD employee told Wall Street News that what impressed him the most was the question about weekends, where Huang Zheng stated, "A colleague asked, when can we have weekends off?" "When we become the second in China, we can have weekends off."

PDD is favored by the capital market due to its high employee efficiency. In 2023, PDD's employee count increased to 17,000, generating revenue of 247.6 billion yuan, with each employee supporting at least 14.5 million yuan in revenue per year, which is 7 times that of JD and 3 times that of Alibaba.

In the rise of PDD, AI algorithms have played a very important role. By analyzing user social relationships, geographical locations, historical behaviors, and other data through distributed AI technology, PDD has constructed dynamic user profiles, achieving "personalized product recommendations" to realize its vision of "goods finding people."

In Huang Zheng's vision in 2018, the next stage of PDD is a combination of "Costco" and "Disney" driven by a distributed intelligent agent network (rather than the currently popular centralized super-brain AI systems), integrating high-cost performance products and entertainment.

"It not only efficiently matches information but also continuously simulates the collective emotions of the crowd in the entire space and attempts to adjust the entire space to make the collective experience happier," Huang Zheng said.

Based on this, PDD's AI and algorithm strategy will not blindly chase technological trends but will focus on "cost reduction and efficiency enhancement" and "user experience."

PDD is facing a significant technological revolution, but on one side is Huang Zheng's "down-to-earth" philosophical constraints, and on the other side is the capital market's strong desire for AI imagination.

Can PDD once again stage a "surprise comeback"? No one knows the answer