Wang Xiaochuan is going to raise money again!
A few days ago, some media reported that Baichuan Intelligence, founded by Wang Xiaochuan, is in the process of a new round of financing of hundreds of millions of dollars. Some investors involved in it revealed that Baichuan Intelligence has been valued at more than 1.8 billion US dollars (about 13 billion yuan) since the beginning of this year. This is just half a year after it completed its last round of financing of $300 million.
However, Baichuan Intelligent responded that the company had not disclosed the latest financing information, and the amount involved in the report was also inaccurate. The refinancing rumors of Baichuan Intelligence are just a microcosm of the continued popularity of the AI large model track since 2023. Continuing this popularity, entering 2024, although the attitude of capital towards AI large models has been differentiated, the track still attracts attention from all walks of life.
The financing of domestic AI large models has repeatedly set new records
Baichuan Intelligence is a Chinese company that develops and provides general artificial intelligence services, aiming to build a Chinese version of OpenAI's basic large model and disruptive upper-layer applications, founded by Wang Xiaochuan, a former CEO of Sogou. Regardless of whether the rumors are true or not, Baichuan Intelligence is already a star project that is highly concerned by capital in the AI large model track.
In October last year, Baichuan Intelligent announced the completion of the A1 round of strategic financing of 300 million US dollars (equivalent to about 2.194 billion yuan), which was participated by Alibaba, Tencent, Xiaomi and other technology companies and a number of top investment institutions. The single financing amount of 300 million US dollars not only allowed Baichuan Intelligent to refresh the record of 2.065 billion yuan outside of Meituan's acquisition of Lightyear last year, but also put it in the first echelon of the financing record of the large model track, and also made it successfully promoted to the list of unicorns.
Entering 2024, the investment and financing of the domestic AI large model track will be "king bomb" again. In February this year, the AI startup "Dark Side of the Moon", which has only been founded for a few months, completed a new round of financing of more than $1 billion, with investors including Sequoia China, Xiaohongshu, Meituan, Alibaba, and old shareholders following the investment. It has already received more than $200 million in financing in 2023. So far, the dark side of the moon has been valued at about 2.5 billion US dollars, making it one of the leading companies in the field of large models in China. It is worth noting that as soon as Kimi, a large-scale model product on the dark side of the moon, was launched, the traffic increased sharply, and there was also a phenomenon that Kimi's concept stocks once skyrocketed. This financing of the dark side of this month is also the largest single round of financing received by a domestic AI large model company since ChatGPT set off a global wave.
Since then, almost every month, AI large-scale model companies have announced the completion of financing. In March, Zhipu AI completed a new round of financing, with the participation of Beijing Artificial Intelligence Industry Investment Fund. It is reported that the valuation of Zhipu AI has exceeded 10 billion yuan, and it is currently one of the fastest startups in China with a valuation of more than 10 billion yuan.
Subsequently, in April, Facewall Intelligence also completed a new round of financing of hundreds of millions of yuan, which was led by Primavera Venture Capital and Huawei Hubble, and followed by Beijing Artificial Intelligence Industry Investment Fund. Founded in August 2022, the core members of the team are from the Natural Language Processing and Social and Humanistic Computing Laboratory (THUNLP) of Tsinghua University, which has also supported the "half of the country" since the wave of large-scale model entrepreneurship last year, and has run out of artificial intelligence companies such as Shengshu Technology and Shenyan Technology.
In addition, the new generation of cognitive and decision-making artificial intelligence enterprise incubated by Lenovo Group, Nordi Intelligence, has also officially completed the Pre-A round of financing, led by Wu Yuefeng Kechuang and followed by Trilobite Venture Capital and Qingzhi Capital, with additional investment from Lenovo Group, with a total financing amount of nearly 100 million yuan. Nordic Intelligence focuses on the industry model and application of "AI + manufacturing", builds a neural evolution system for enterprises, achieves accurate and continuous optimization, and then realizes efficient operation.
The attitude of capital is differentiated
Some investors suggested taking the opportunity to exit
Although the financing amount of AI large model companies is higher than the other, from the perspective of the global investment and financing situation of this track, there has been a structural cooling.
According to a report prepared by PitchBook, venture capital firms invested $25.87 billion in AI startups globally in the first quarter of 2024, up from $21.69 billion in the first quarter of 2023. However, the investment in the first quarter of 2024 involved only 1,545 transactions, compared to 1,909 in the same period in 2023. Meanwhile, M&A transactions slowed from 195 in Q1 2023 to 176 in Q1 2024.
John-David Lovelock, an analyst at the Institute for Human-Centered Artificial Intelligence at Stanford University, believes that from a global perspective, the number of billions of dollars of investment has slowed down and is almost over, and the current hot money is pouring into a new direction - AI applications, so if 2023 is the first year of the outbreak of large models, 2024 will be the year of the explosion of large models into the application, and capital will undoubtedly pay more attention to how to apply large models.
Therefore, many domestic investors have expressed doubts about the investment model. First of all, Zhu Xiaohu, a partner of GSR Ventures, bluntly said in an interview with the media that he was not optimistic about the development of domestic large models. The reason is that the large model technology has no scenes, no data, and the valuation is still very high, and the investment win rate is very low.
In addition, Li Gangqiang, a senior investor, also agreed with Zhu Xiaohu's point of view, "If I had to choose, I would definitely focus on the application of large models." He also suggested that early-stage investors should take the opportunity to exit the general model project, and the window of time left for early investors will not be too long. The reasons are: first, the giants advance into the large model, and the ordinary entrepreneur is only the denominator; second, the domestic general large model technology can not catch up with foreign countries, once the foreign large model is open source, the value of the domestic large model will be wiped out; third, the large model is extremely money-burning, and in the domestic capital downturn, massive financing will become more and more difficult; fourth, although the valuation of the large model has risen through continuous financing, the return on investors is very low. In Li Gangqiang's view, the current capital cycle curve in the field of large models will soon reach the right side of the parabola, and there will be no chance if you don't exit.
However, some "technical" investors believe that it will still take time for large models to polish the underlying capabilities, and it is still too early to talk about applications. At the 2024 NVIDIA GTC, Liu Wei, partner of Broadband Capital, said when sharing the entrepreneurship and investment opportunities in the AI industry in the Chinese market, he is personally very optimistic about the development of AI as a whole, and from the perspective of investment, he does not believe that AI companies should take commercialization and profit as a major goal in 2024. Even in the U.S. market, where AI is developing faster, there are some seemingly successful application companies, but they have not been able to find an effective business model. In his opinion, the current large-scale model capabilities are still rapidly iterating, and when the large-scale model capabilities are limited, it is still too early to talk about application scenarios, let alone commercialization. It is natural to succeed first in large model capabilities, then on the application side, and finally commercialization.
The most beautiful April day in the world, living up to the spring and the times. In this season of growth, we ushered in the grand event of the domestic venture capital industry - the 2024 Securities Times "Looking for the 'Golden Eagle' of Venture Capital and Discovering the 'New Seedlings' of Enterprises" plan was grandly relaunched.
Click on the picture below to scan the code, or "Read the original article" in the lower left corner to register!
![Xiaohongshu made the agents quarrel! Jointly with Fudan to launch an exclusive group chat tool for large models[fig]](data:image/gif;base64,R0lGODlhAQABAIAAAP///wAAACwAAAAAAQABAAACAkQBADs=)