China’s Tencent Holdings has officially exited its investments in e-commerce giant JD.com and food delivery firm Meituan, as the tech giant pulls back from the cash-burning internet economy to focus on artificial intelligence (AI).
Co-founder and CEO Pony Ma said at the company’s annual staff meeting on January 26 that “a steady and pragmatic approach remains one of Tencent’s core strategies” and that “AI is now the only area worth investing”.
Ma also disclosed at the staff meeting that Tencent’s president Martin Lau has ceased to be on the boards of both companies.
Tencent’s exit from the two internet economy leaders comes after a year of cutthroat competition in the on-demand delivery sector. Known as the year of “a brutal price war”, 2025 saw JD.com enter the market with billions of Chinese yuan pledged in consumer subsidies to challenge the established player Meituan, as well as Alibaba Group, which also upped its game last year by integrating its food delivery app Ele.me into Taobao Flash Buy.
Tencent became a shareholder of the then group-buying site Meituan in 2014, one year before Meituan merged with domestic restaurant review platform Dazhong Dianping to create a dominant online-to-offline (O2O) service giant in China. It later served as a cornerstone investor of Meituan’s $4.2 billion Hong Kong initial public offering (IPO) in 2018. Its shareholding in Meituan was at one point close to 20%.
As for JD.com, Tencent first acquired a 15% stake in the company for about $214.7 million in early 2014—right before JD.com’s $1.78 billion Nasdaq IPO—to aid its e-commerce fight against Alibaba. It later increased its stake in the company to almost 20%.
Tencent’s divestments from both companies had started around 2021-22 through billion-dollar dividend distributions to shareholders, in part due to the Chinese government’s sweeping antitrust probes and regulatory crackdown on tech giants.
Strong IPO markets reward Tencent’s AI bets
Tencent, famed for its ubiquitous messaging service and online payment app WeChat, has been quietly restructuring its investment exposure. But this transition has become public only now, as a strong IPO momentum offers the potential of high rewards from its AI bets.
Zhipu AI, in which Tencent first invested in 2023, completed the first Hong Kong IPO by a Chinese large language model (LLM) company in January, raising over $550 million. Its peer MiniMax, whose Series A round in 2023 was led by Tencent, raised $620 million in its Hong Kong IPO a day later.
Besides the two completed IPOs, Tencent-backed CXMT Corporation, the parent of Chinese memory chip giant ChangXin Memory Technologies (CXMT), is poised for a blockbuster IPO of up to $4.2 billion on Shanghai’s Nasdaq-style STAR Market, which is set to become the board’s second largest since its launch in 2019.
AI chip unicorn Enflame Technology, which counts Tencent as a shareholder with a significant near-20% stake, is also moving towards a STAR Market IPO, targeting to raise over $830 million.
As Ma put it, “AI is now the only area worth investing,” Tencent is clearly shifting away from its internet-era playbook of funding platform businesses heavily reliant on a substantial market share and strong user traffic, to these capital-intensive, high-tech-barrier AI infrastructure opportunities.
While some investors are concerned that Tencent’s large AI model lags those developed by Alibaba and ByteDance, we expect Tencent to increase its capex in fiscal year 2026, said analysts from financial services firm CGS International. “China’s AI players will speed up AI model upgrades and monetisation this year, with large players competing to become the primary traffic entrance for AI agents,” they shared in an emailed note on February 3.
Tencent’s determination in AI is also reflected in a few recent recruitment successes. In December 2025, Tencent appointed 27-year-old former OpenAI researcher Vinces Yao Shunyu as its chief AI scientist. Last week, the tech giant also hired Pang Tianyu, previously a senior research scientist at Sea AI Lab in Singapore, for its Hunyuan Multimodal Department, to lead the exploration of reinforcement learning algorithms.
