In the third quarter, the number of AI financing increased by nearly 100% year-on-year; the Yangtze River Delta urban agglomeration showed a significant growth rate; Changsha unexpectedly made it into the top ten in the country.
The artificial intelligence industry is one of the most prominent focuses in the current venture capital circle. We should not only pay attention to the industry's development trends and technological hotspots but also understand the capital flow and changes in the AI circle.
In the just - ended third quarter, how did the overall investment data of the AI industry perform? Without further ado, let's start with a few statistical charts.
I. Overall financing scale: The "scissors gap" between the sharp quarter - on - quarter increase in the number of events and the slight increase in the amount in Q3
From the first quarter of 2024 to the third quarter of 2025, the new financing in the domestic AI industry showed the characteristics of "continuous expansion in the number of events and fluctuating and differentiated amounts".
Note: In this article, the artificial intelligence industry adopts a relatively broad definition, covering tracks such as AI chips, autonomous driving, and robotics.
According to IT Juzi data, from Q1 2024 to Q3 2025, the number of new financing events in the domestic AI industry achieved continuous and substantial leaps. By Q2 2025, it almost doubled from 188. In Q3 2025, it reached 435, a quarter - on - quarter increase of 20.8% and a year - on - year increase of 99%.
It can be said that compared with last year, the artificial intelligence industry once again witnessed a new round of financing boom in 2025.
Judging from the estimated financing amount, the overall impression of the amount curve is a fluctuating and growing change.
Due to factors such as some enterprises not disclosing their financing amounts and the influence of occasional large - scale financings, the data trend of the total financing amount shows a certain degree of uncontrollability and may have deviations. Therefore, the "number of events" can more truly reflect the industry's investment activity, while the financing amount mainly represents the intensity of capital investment. The data is for reference only.
In the third quarter of this year, the number of AI financing events increased by 20.8% quarter - on - quarter, but the estimated total amount increased by only 3%, with an overall scale of 37.037 billion yuan. The fundamental reason for this "scissors gap" lies in the "structural imbalance in the investment stage".
The number of early - stage investment events increased significantly, from 144 in Q2 2025 to 214, an increase of 49%. In contrast, the mid - and late - stage investments basically stagnated and even decreased. Since the single - round amount of early - stage investments is generally low, even with a sharp increase in the number of events, it is difficult to drive a significant increase in the total amount, resulting in only a slight increase in the overall financing amount in Q3.
II. Investment stage distribution: The early stage becomes the "absolute main force"
Note: Early - stage investments include seed rounds, angel rounds, and Pre - A rounds. Mid - stage investments include A rounds, A+ rounds, Pre - B rounds, B rounds, and B+ rounds. Late - stage investments include C rounds and subsequent stages up to the Pre - IPO stage before listing.
As shown in the IT Juzi data in the above figure, in the past two years, the number of early - stage investment events in the field of artificial intelligence has increased significantly.
Looking at the timeline, since Q2 2024, early - stage investments have increased quarter - on - quarter for six consecutive quarters (14%→3%→22%→22%→2%→49%), reflecting the continuous increase in the importance that capital attaches to the "early - stage layout" of AI.
In Q3 2025, early - stage investments increased by 49% quarter - on - quarter, setting the highest growth rate in the past seven quarters. There are two core reasons behind this growth:
Firstly, AI technology has expanded in various application levels and segmented tracks, such as embodied intelligence, AI healthcare, and intelligent driving. The number of new startups has increased significantly, and capital needs to "cast a wide net" to seize the technological opportunity.
According to IT Juzi statistics, among the AI enterprises that received investment in the third quarter, 73 were established in 2025 and 97 in 2024, accounting for a total of 39%.
Secondly, early - stage investments have relatively controllable risks (low single - round amount and small trial - and - error cost), making them a "safe option" for capital layout.
The mid - stage investment increased by 34% quarter - on - quarter in Q2 2025, but only by 5% in Q3 2025. The slowdown in growth means that capital may pay more attention to enterprises with "clear commercialization implementation". Enterprises lacking implementation capabilities will find it difficult to obtain subsequent financing.
After the number of late - stage investments (C rounds and later) increased to 43 in Q2 2025, it dropped to 34 in Q3 2025, and the overall scale has always been lower than that of the early and mid - stages. The core reason is the "scarcity of late - stage targets" in the AI industry. At this stage, capital tends to concentrate on investing in leading enterprises with "high technological barriers and stable market shares".
Strategic investment, with industrial capital as the main player, is increasingly becoming an "important supplementary force" in AI industry investment.
III. City distribution: "Beijing, Shanghai, Shenzhen, and Hangzhou" firmly remain in the first echelon, and new first - tier cities "break through with differentiation"
From the perspective of cities, which cities' artificial intelligence industries were favored by investors in the third quarter of this year?
The ranking of cities in AI financing not only reflects the differences in urban industrial bases but also the direction of regional policy and resource allocation.
"Beijing, Shanghai, and Shenzhen" in the first echelon monopolize resources with different focuses
Beijing is the top destination for AI investment in the country, but its year - on - year growth rate is lower than the overall level. Among them, "Haidian District" is at the core. AI investment accounts for 64.8% of the city's financing transactions and 74.1% of the financing amount. It has gathered AI star enterprises such as Zhipu, a large - model enterprise, and Xingdong Jiyuan, an embodied - intelligence company. Relying on the location advantage of Zhongguancun and the technological and talent advantages of top universities such as Tsinghua and Peking University, it has become a "high - tech R & D center for AI". It is worth mentioning that in Q3, Chaoyang District saw a more than 130% growth rate in AI financing events.
Shanghai's total number of AI events has a higher year - on - year growth rate (109%) than Beijing (61%). Especially in "Xuhui District", AI enterprises performed outstandingly in the third quarter, with a year - on - year increase of 300% in financing events. For example, Minimax, a large - model enterprise, and Tashizhihang, an embodied - intelligence enterprise, both received high - amount financing, mainly relying on the policy support of the Xuhui Riverside AI Industrial Park. Pudong New Area focuses on AI computing power (such as Lightelligence) and intelligent driving, with a more solid industrial foundation.
Shenzhen ranks third in the country in terms of the number of new AI financing transactions, and its year - on - year growth is more than double. Different from the "soft AI" (large models, computing power) pattern in Beijing and Shanghai, Shenzhen's AI financing projects are more inclined to "hardware", such as Zhongqing Robot and Zibianliang Robot. Relying on the supply - chain resources of electronic manufacturing giants such as Foxconn and Huawei, Shenzhen has become an "AI + manufacturing implementation demonstration center".
Based on the Alibaba ecosystem, Hangzhou's AI projects are mostly "bound to application scenarios", such as the embodied - intelligence robot of Qianxun Intelligence and the AI Agent of Better Yeah AI. With strong commercialization implementation capabilities, Hangzhou ranks fourth in the country in terms of the number of new AI financing events and the scale of financing amount.
New first - tier cities "break through with differentiation", and the regional cooperation in the Yangtze River Delta accelerates
Chengdu, Nanjing, and Changsha are in the category of "high growth rate and low amount" with a small base. They mainly focus on early - stage project investments. Their core advantage is "low cost + policy subsidies", including tax exemptions and rewards for AI startups and support programs for university entrepreneurial talents, which attract capital for "early - stage trial - and - error layout", but they have not yet formed a mature industrial cluster.
In addition, cities in the Yangtze River Delta such as Hefei, Suzhou, and Nanjing have shown significant growth rates. In particular, the number of new AI financing transactions in Suzhou's artificial intelligence industry increased by 283% year - on - year, covering AI medicine, AI chips, AI applications, etc. This also reflects the prominent "synergistic effect of the Yangtze River Delta AI industry". The agglomeration and spill - over (relocation or new establishment) of Shanghai's artificial intelligence industrial cluster enable the major cities around Shanghai to undertake the manufacturing and implementation links in the AI industrial chain, forming a regional closed - loop of "R & D - production - application".
Mid - western cities such as Chengdu and Changsha rely on their local industrial bases (such as Chengdu's electronic information and Changsha's engineering machinery) to explore a differentiated path of "AI + traditional industries" to avoid direct competition with Beijing, Shanghai, and Shenzhen.
IV. Three core findings of China's AI investment in Q3 2025
The investment enthusiasm in the industry has increased:
In Q3 2025, the number of domestic AI industry financing events doubled year - on - year, confirming that AI remains a focus in the venture capital circle. Moreover, the investment enthusiasm this year has been higher than that of the same period last year for three consecutive quarters.
Capital preferences vary at different stages:
In the third quarter, early - stage AI investments increased significantly, while mid - and late - stage investments remained flat. This reflects that a large number of new AI startup projects and technological segmented directions are emerging. Capital has a higher risk preference for early - stage layout, with the core strategy of seeking "breadth", while for mid - and late - stage layout, it pursues "precision".
The regional pattern is solidified, and new first - tier cities need to "differentiate":
Beijing, Shanghai, Shenzhen, and Hangzhou, relying on their industrial ecosystems, technological talents, and capital resources, almost monopolize the resources for AI entrepreneurship and financing. New first - tier cities such as Chengdu, Suzhou, and Nanjing need to rely on policy and state - owned capital support and their local industrial bases (such as Suzhou's automobile industry and Chengdu's electronic information industry) to explore a differentiated path of "AI + traditional industries" to break through the pattern limitations.
This article is from the WeChat public account "IT Juzi", author: Wu Meimei. Republished by 36Kr with permission.