Alibaba, Meituan, and Tencent's unrealized gains from these three AI companies have already exceeded 100 billion yuan, making huge profits on investments, but compared to the market value evaporated by these three companies, it's a huge loss.

But when it comes to investing in AI, the styles of the three are completely different. Meituan invests the least but earns the most, Alibaba casts a wide net with some investments in computing power, and Tencent is in the middle.
1. First, look at Zhipu: 45x for Meituan, crushing the competition.
Zhipu's current market value is about 350 billion RMB (Hong Kong stock 02513, briefly broke through 1,000 HKD during intraday in early May, with a total market cap exceeding 400 billion HKD), Zhipu's GLM5.1 ranks in the top 3 globally for large models, and their coding package sells like crazy every day.
Meituan was the earliest to bet. In March 2023, during Series B2, they invested 300 million RMB directly, with a post-investment valuation of 3.2 billion RMB, holding over 10%. After that, Meituan did not add more, and through subsequent rounds of financing and IPO dilution, they still hold 3.91%, corresponding to a market value of 13.7 billion RMB.
300 million RMB turned into 13.7 billion RMB, a net profit of 13.4 billion RMB, a 45x return. This figure is top-tier in the entire Chinese primary market.
Tencent came a bit later. In August 2024, during Series B4, they invested 200 million RMB, with a post-investment valuation of 7.2 billion RMB, holding 2.7%. After dilution, they hold 1.58%, worth 5.5 billion RMB, a 27x return.
Not bad, but compared to Meituan, which invested over a year earlier, their return was cut in half.
Alibaba's path is the most convoluted. Ant Group's Shanghai Yunyu first subscribed to 666,700 RMB of registered capital in Zhipu during Series B3 with 150 million RMB, and after IPO, they held 1.54%, worth 5.4 billion RMB, a 33x return.
Alibaba itself isn't bad; Ant Group, through Shanghai Yunyu and Shanghai Feiya, holds a total of 3.66%, worth 12.8 billion RMB, with a cost of 490 million RMB (after deducting 110 million RMB transferred to Alibaba), a 25x return.
The Alibaba ecosystem (Alibaba + Ant) holds a total of 5.2% in Zhipu, with an unrealized gain of about 17.5 billion RMB.
2. MiniMax: Alibaba is heavily invested, Tencent co-invests, Meituan is absent.
Alibaba has placed a heavy bet on MiniMax. Through Alisoft, they hold 12.52%, which, based on the current market value, is about 3.7 billion USD.
Alibaba participated in the B round and cornerstone round, but MiniMax did not disclose Alibaba's exact investment amount. Based on rough estimates, the cost is about 600 million USD, with an appreciation of about 5.2 times, earning 3.1 billion USD (667k RMB).
Tencent owns 2.37%, valued at about 700 million USD, participating in earlier rounds than Alibaba, with an estimated cost of around 100 million USD, appreciating roughly 6 times.
Although the absolute amount is less than Alibaba's, their cost control is better, and the multiple is higher.
Meituan did not participate in MiniMax's investment.
3. Kimi: Alibaba might be the biggest winner, Meituan also heavily invested.
Kimi's latest round, a $2 billion financing, just completed, with a post-money valuation exceeding $20 billion (~RMB 1.4 trillion), led by Meituan Longzhu.
Alibaba was Kimi's most important early financial investor. In 2024, they bought about 36% equity for $800 million, some settled with Alibaba Cloud computing power, roughly $600 million in cash.
In February 2026, they participated in another $700 million financing, with undisclosed specifics.
Based on 36% equity, the current valuation is about $7.2 billion, with a book return of about 9 times. But this multiple depends on Alibaba not being heavily diluted in subsequent financings and the $20 billion valuation holding up—both are uncertain.
Tencent participated in a $300 million financing at a valuation of $3.3 billion in August 2024, and in February 2026, they co-led a $700 million financing, but the amounts are undisclosed, so precise unrealized gains can't be calculated.
Following Tencent's usual style, they are likely co-investors rather than heavy shareholders.
Meituan's side involves two investments. Wang Huiwen's personal total investment is about $70 million (roughly RMB 490 million), with a current estimated return of about 5x—an early investment that has earned decent returns.
Meituan Longzhu led this round of $2 billion financing, with a single investment exceeding $200 million. After the investment, the valuation didn't increase much, so there is currently no unrealized gain.
Some interesting points:
1. Meituan's return multiple crushes others. Investing 300 million RMB, exiting at 13.4 billion RMB, a 45x return. The reason is simple: invested early, exited early. They entered during Series B2.
2. Alibaba invested in Zhipu, MiniMax, and Kimi, with the largest total unrealized gains, each with significant investment amounts, and a large portion invested in computing power.
3. Tencent is the most balanced. Zhipu 27x, MiniMax 6x, overall return 11x, Kimi estimated at 6x.
4. Kimi might be Alibaba's biggest single-return source—36% equity valued at $20 billion, worth $7.2 billion. If the final IPO pricing is higher, this figure could expand further.
These three AI companies' stock prices are highly volatile. Zhipu's stock price soared from 131 HKD on listing day to over 840 HKD, MiniMax from 165 HKD to over 1,200 HKD, but they also experienced sharp corrections. Before these unrealized gains are realized, they are all paper wealth.
Another point worth mentioning: the investment logic of these three giants in AI isn't purely financial. Alibaba needs computing power clients, Tencent needs ecosystem positioning, Meituan needs technological barriers—all are investing for different reasons but with the common goal of avoiding replacement.