Cao Xi is going to raise US dollar funds again, and the target fundraising amount this time is 500 million US dollars.

Over the past few years, the narrative that "USD fundraising is back" has been touted every few days, just like the claim that "consumer investment is viable again". But until recently, all signs indicate that this time it seems truly different.

The most intuitive indicator is that GPs are almost collectively active in the USD fundraising market. Since the second half of last year, a number of well-recognized VCs have repeatedly appeared in foreign media reports, including HSG, IDG, Matrix Partners, Source Code Capital, Shunwei Capital, Vision Capital, InnoAngel, ZhenFund, Qiming, etc., all of which have been exposed to have planned or launched new rounds of USD fundraising. According to my side-line understanding, 5Y Capital and Gaorong Capital are also likely on this list.

According to reports, IDG is reportedly preparing a growth fund with a target size of about 1.2 billion US dollars; HSG plans to launch a new VC fund while preparing a new RMB fund; the new fund of Shine Capital focuses on AI, with a target size slightly higher than 200 million US dollars.

Another clear case is Monolith, which has strong momentum. A week ago, Monolith was reported by foreign media to have launched a new USD fund with a target size of about 500 million US dollars, focusing on AI, almost doubling the size of the previous $288 million fund.

Just today, Bloomberg reported that Boyu Capital, GL Ventures and Chunhua Capital have also launched fundraising in recent months.

Some institutions have already handed in their results this year. At the beginning of the year, Source Code Lv completed its first $150 million fund with over-subscription; in March, Joe Tsai's Blue Pool Capital raised its first $1 billion private equity fund; in April, BlueRun Ventures announced the completion of fundraising for its fourth dual-currency fund, totaling 560 million US dollars; two days later, Lightspeed China Partners announced the completion of a new dual-currency fund, of which the early-stage USD fund amounted to 460 million US dollars.

In May, BAI Capital completed the first close of its new USD fund at 600 million US dollars, with a target of 800 million US dollars, and this fund was launched only at the beginning of the year. In July, Black Ant Capital completed its new USD fund and achieved over-subscription. According to foreign media reports, Qiming and ZhenFund have also completed new rounds of USD funds; ChinaVenture has exclusively learned that the new USD fund of YoVentures has been over-subscribed and fully closed.

According to additional side-line information, Source Code Lv, which just closed its first $150 million fund in January, will also launch a new round of USD fundraising within this year. In the past year, Source Code Lv has been actively making investments in AI and embodied intelligence, so this move is easy to understand.

If we only look at these figures, the situation seems quite optimistic, but the real answer lies beneath the surface. After talking with several GPs, IRs and LPs who are in the process of fundraising or have just completed fundraising, I think a more accurate description is: the attention of capital is back.

As for whether the capital itself has returned? That is probably another matter.

From "Holding One's Tongue in Fear" to "Revisit"

First of all, this is a collective restart of USD VCs after three or four years.

In the past, the fundraising rhythm of USD funds was roughly one new fund every two years, and the last wave of concentrated fundraising was in 2021 and 2022. Launching a new fund last year or this year means the interval has been extended to four or five years. There are naturally many reasons: changes in the macro and policy environment, fundraising uncertainty, asset shortage and exit difficulties, all of which have significantly slowed down the investment pace of GPs. It was not until 2025 that large AI models brought a new round of entrepreneurial opportunities, coupled with the fact that old funds gradually entered the late stage of investment. Multiple factors together made 2026 a year of concentrated fundraising.

The restart is a collective action, and over-subscription only occurs in individual cases. Today's over-subscription generally has only two scenarios: the privilege of top-tier institutions, or over-subscription on the premise of lowering the original target.

A USD IR told me that their institution launched USD fundraising at the coldest market trough two years ago, and the original target was reduced by about 30% after LPs repeatedly persuaded them to lower expectations. As the primary market warmed up this year, and with several impressive IPOs delivered, the fund was finally completed with over-subscription, but the actual size was basically the same as the original target before the reduction.

The more common fact is that USD VCs have actively lowered their expectations. At the fundraising peak in 2021 and 2022, $500 million was even a regular size, and $1 billion was not rare at all. The performance back then was: Source Code Capital raised $1 billion, 5Y Capital raised a total of $1.56 billion for two funds, Qiming's eighth USD fund reached $2.5 billion, and for top-tier institutions, HSG raised a total of about $9 billion for four funds in 2022.

Four or five years later, a USD IR told me that the target size of leading institutions has generally dropped to $500 million to $1 billion, and most funds are concentrated in the range of $200 million to $300 million. In this case, if Monolith finally completes $500 million in fundraising, it will be a rare counter-trend expansion, with a size almost close to that of leading institutions.

Of course, the main theme is contraction. Preqin data shows that in 2022, a total of 1105 USD funds focusing on the Chinese market raised a total of 1.5 trillion US dollars, while in 2025, only 97 funds raised 13.6 billion US dollars, less than one-tenth of the size in 2022.

The fundraising cycle has also been extended. In the 2021 and 2022 wave, it was not uncommon to complete fundraising in three or four months. According to the experience of USD IRs, most funds now generally need half a year or even longer.

The market is fragmented in the sense that almost every participant in USD fundraising I talked to said that since the fourth quarter of last year, the number of USD LPs taking the initiative to interact has increased significantly.

Long Yu, Founding and Managing Partner of BAI Capital, told me that the number of LPs taking the initiative to communicate has increased significantly since this year, and their sources are more diverse. In addition to non-US LPs from the Middle East and Europe, Southeast Asian sovereign funds are also re-evaluating China, and even traditional LPs that have allocated to China for a long time have begun to "revisit". Another IR told me that family office LPs, who were once absent and have the most sensitive market sense, have also reappeared.

They are enthusiastic in attitude and optimistic about the market. A USD IR mentioned a celebration banquet where a RMB-denominated project was listed on the A-share market, and USD LPs also showed up to celebrate. This cross-currency celebration scene would have been hard to imagine a few years ago.

This decent and optimistic interaction easily creates the illusion of "reconciliation". In fact, most interactions are still at the stage of discussion and evaluation. Long Yu used an accurate description: the attitude of USD LPs towards China has only changed from "holding one's tongue in fear" in the past to "revisiting now".

The reason why fundraising is concentrated in 2026 is not entirely due to the opening of the secondary market or the boom of the Hong Kong stock market.

The reaction of USD LPs is often one beat slower than local investors. The turning point really started with the explosion of DeepSeek in February 2025, and a series of technological breakthroughs continued to loosen the attitude of USD LPs. Chen Weiguang, Managing Partner of BlueRun Ventures, clearly felt this when he announced the new fund. A USD IR told me that overseas institutional LPs usually make their budgets for the next year at the end of each year. After seeing DeepSeek in 2025, they could only conduct research and observation first, and would consider including Chinese GPs again when making the 2026 budget at the end of the year.

In other words, the collective revisit in 2026 is a new procedural action for USD LPs after a year of research and consideration.

From "Why China" to "Which China"

In the past three years, GPs often needed to answer the question "Why China". This year, this question has disappeared. A USD IR said that in previous years, LPs cared about macro issues, but now the question has become "What in China is worth investing in", with increasingly specific granularity, such as which AI sub-sectors to focus on, whether there are still AI application opportunities in China, and whether enterprises must go global.

The industry generally regards the Middle East, Europe and Southeast Asia as the new base of USD LPs, and Hong Kong and Singapore are also very active, but it is hard to say that any party is particularly aggressive in making investments. An IR shared a counter-consensus observation: the industry generally believes that the Middle East is the major LP after the United States, and GPs are flocking to the Middle East for fundraising, but the number of institutions that have actually obtained capital is not as large as imagined.

"Some Middle East LPs will set up return-investment requirements that are not easy to meet, and their decision-making cycle is very long."

Restricted by policies, US LPs have long reduced their allocation to China to a very low level, but they have not completely withdrawn. A small amount of capital will still make conditional investments. A practical approach is to set up parallel funds to isolate US LPs from clearly restricted investment directions.

The preference of capital is also changing. On the surface, the only theme of this round of fundraising is technology, and AI and robots are the core reasons for USD investors to revisit China, which has reached full consensus just like the RMB market. An IR of a non-technology fund has a very clear experience: in previous years, they would at least look at consumer and biotech tracks, "Now, it is even too conservative to describe the market as an 80/20 effect, the 90/10 effect is more accurate."

AI and robot themed funds are particularly popular. Large models are not the only ticket to success, but they are an extremely important one. It is not difficult to find that this batch of VCs that are raising USD funds all have impressive AI investment track records, which constitute the core narrative when pitching to USD LPs.

FOMO (Fear Of Missing Out) in the primary market has never stopped. Everyone resists FOMO, but is in a state of FOMO all the time. But when it comes to investing in large models, FOMO does make sense. A USD IR explained directly: "If you have not invested in leading large model and robot projects, it is very difficult to demonstrate to LPs why you can continue to succeed in AI investment."

When GPs choose to launch fundraising and what chips they have in hand is a very dynamic process. The status of star projects will affect the fundraising window to a certain extent. The best chips in the first half of the year were Zhipu AI and MiniMax, which were listed on the Hong Kong stock market. "It is very easy to tell a good story when you launch fundraising when the stock price performs well." In the second half of the year, DeepSeek, which opened financing, and Kimi, which became popular again relying on K3, have stronger persuasion. LPs may even ask GPs that have invested in Zhipu AI and MiniMax: why didn't you choose Kimi?

An interesting detail is that last year, Monolith was labeled as "DeepSeek rival's backer" in foreign media reports, but this year it became "Kimi's backer", which shows the rising momentum of Kimi. Of course, Monolith has successfully bet on both Kimi and DeepSeek, and IDG is the other institution that has invested in both.

If the story stops here, it will be somewhat disappointing. It seems that USD LPs have switched from a narrow value system to another even more cruel and single evaluation system.

There are two different examples, such as Black Ant Capital. Consumption has become a marginal theme in the primary market, but it is understood that Black Ant Capital's new USD fund was raised very smoothly and achieved over-subscription, purely relying on its strong performance. The confidence comes from the fact that Black Ant Capital has continuously invested in several hit projects such as Pop Mart, Laopu Gold and Mingmang Mangmang in the past few years.

At the end of the day, you have to either catch the most scarce AI assets, or deliver strong returns, you must meet at least one of the two requirements.

Another example is BAI Capital. BAI completed the first close of $600 million in less than half a year. The new fund does not put all its chips on a single AI theme. It remains a growth fund that starts from Asia to seek global opportunities, with relatively diversified industry coverage. In today's market, this strategy even seems a little "not trendy enough".

Of course, BAI is not a completely traditional sample of USD fundraising. BAI was formerly the Bertelsmann Asia Investments Fund under the Bertelsmann Group. It became independent in 2022 and completed its first market-oriented fundraising. Its LPs have expanded to global sovereign funds, insurance capital, internet giants and FOFs, and its fundraising experience is consistent with the market situation.

Long Yu attributes this round of changes more to the re-comparison of global assets: China's relative attractiveness is rising. On the one hand, US AI investment is highly concentrated, with valuations and capital needs continuing to rise; globally, Europe is facing weak growth, and emerging markets such as India have not yet formed a market size and industrial depth comparable to China. When global capital reallocates, China is still an important market that cannot be ignored outside the United States.

In her view, AI is more like a powerful catalyst, but the underlying driving force is still the comparative advantage of China in the global asset system.

In Long Yu's view, today's USD LPs can hardly be convinced by a simple story. Smooth fundraising does not mean that capital has "systematically returned". She remains cautious about this round of recovery, and prefers to regard it as a phased heat rather than a complete reversal of the trend. Even if fundraising is going smoothly at the moment, she still judges internally based on the scenario that "the market may cool down again in a year or so".

In conclusion, everyone is very clear that the old USD era will not come back. USD fundraising is warming up, but the 2026 USD funds are no longer the same as the 2021 USD funds.

LPs have changed, and the themes have also changed. The previous "China growth story" has become the "China technology story". The last round of USD VCs could tell a broad "China growth story": consumption, internet, healthcare and technology could all be part of China Beta. Today, what USD LPs buy is more like China Alpha — AI, robots, and a small number of technology assets with global competitiveness.

GPs have also changed. The most popular institutions in the last generation were all-powerful super platforms. This round, funds are getting smaller, themes are getting narrower, and LPs are more demanding. The institutions that can truly achieve counter-trend expansion may instead be boutique funds like Monolith, which have a lighter organization, more focused strategy, and have caught scarce assets.

The real change can probably be described as: USD investors are willing to discuss China again, but the previous allocation logic, power structure, game rules and mutual relationship have all changed.

This article is from the WeChat official account "ChinaVenture", author: Cao Weiyu, published with authorization from 36Kr.