When LP uses Doubao to teach me investing: A private equity GP's career transition autobiography

Original | Odaily Planet Daily (@OdailyChina)

Author | Golem (@web3_golem)

When LPs learn to use AI, the days for small private equity fund managers are becoming even harder.

Two Dog (@ryansoon777) was still a general partner (GP) of a small offshore private equity dollar fund focused on U.S. stocks in China a few years ago, but after the year he left to join an AI startup.

"It's already difficult for small private equity funds to raise capital, and with AI becoming widespread, many investors (LPs) would rather let Doubao assist in stock trading than give us their money."

Two Dog said that his decision to switch careers was largely because he saw the subtle impact of AI on the relationship between LPs and GPs. As information and analytical capabilities are seemingly leveled by AI, LPs are more likely to question the professional judgment of GPs, which could increase friction between the two sides, and in severe cases, even lead to capital withdrawal or exit.

Small dollar private equity funds, which are already struggling

The private equity dollar fund where Two Dog previously worked was not poorly operated, with assets under management reaching several tens of millions of dollars, mainly investing in highly liquid U.S. stocks, with a small portion in crypto asset management. Its annualized return over the past three years has far exceeded the Nasdaq.

Logically, with excellent performance and increased overseas wealth management demand in recent years, fundraising should not be difficult, but Two Dog revealed that, in fact, small dollar funds like theirs are almost impossible to attract institutional LPs.

Currently, top domestic dollar private equity funds with hundreds of millions of dollars (such as Jinglin, Hillhouse, Boyu, etc.) mainly adopt a "offshore + onshore" combined structure, where the fund entity stays in Cayman, often registered as a Cayman exempt company or Cayman SPC, with the management entity located in Hong Kong or Singapore.

However, in recent years, due to regulatory and fundraising environment changes, more private dollar funds are adopting onshore structures based solely on Hong Kong LPFs or Singapore VCCs.

And the small dollar private equity funds like the one Two Dog joined still use the most "primitive" dollar fund structure, namely Cayman SPC + BVI (British Virgin Islands) fund manager structure.

A common saying in the fund industry is that LPs decide the structure. The reason why top domestic dollar private equity funds still cling to "Cayman" is partly because their overseas LPs include American university endowments, Middle Eastern sovereign wealth funds, and large European family offices. These top-tier "old money" investors have been familiar with Cayman structures for decades, and continuing to use this setup helps reduce communication and trust costs.

But domestic small dollar private equity funds based in the Cayman are unlikely to attract these top international funds, as their LP sources are mainly in Asia. This puts them in an awkward position.

From an Asian perspective, the main backers of dollar private equity funds are private banks, mainland Chinese (outbound capital), local Hong Kong family offices, and Southeast Asian wealthy individuals.

Even for small dollar funds of similar size, these circles tend to have a natural affinity and sense of security towards Hong Kong or Singapore structures, so they prefer investing in Hong Kong LPFs or Singapore VCCs rather than Cayman SPCs.

Besides fund structure and size limiting fundraising channels, differences in investment strategies also make it harder for Two Dog and others to raise funds.

In private equity investment strategies, there are mainly subjective strategies and quantitative strategies. Subjective strategies rely on GPs' research, experience, and judgment to decide what to buy or sell, with the core profit coming from the fund manager’s market insight; quantitative strategies involve coding investment logic into mathematical models and programs, which then automatically or semi-automatically execute high-frequency trades, with profits driven by the statistical regularities used in the models.

"Currently, funds using quantitative strategies are easier to raise capital than those using subjective strategies, especially after AI empowerment, LPs are more convinced by quant strategies," Two Dog said, especially after DeepSeek (Odaily note: incubated by the quant fund Fantasia Quant) exploded in popularity last year, boosting market enthusiasm for quant strategies.

Furthermore, the difference between quantitative and subjective strategy funds is that quant strategies can demonstrate data and algorithms to LPs to gain trust. Whether profits or drawdowns, they are within controllable ranges. Excellent quant funds can even serve as fixed-income products; subjective strategies are more abstract, and GPs need to spend more effort to gain LP trust. During large drawdowns, LPs are more likely to question the investment ability of GPs.

Therefore, in summary, the survival space for small dollar private equity funds like the one Two Dog previously worked for has been squeezed by the broader environment, and fundraising is becoming increasingly difficult. And some of the remaining large LPs within these funds are also questioning whether AI’s "investment ability" surpasses that of GPs.

"Complex" LPs

"Previously, LPs would listen to us because we are professionally trained, but now they just give our reports to AI to translate into plain language, then turn around and 'teach' us how to do things," Two Dog said that after AI became widespread, LPs who used to only look at the final results now show much more concern about his investment operations.

He even once had to remove an LP. This was a 50-something entity owner, very "hands-on," who invested about $1 million in the fund where Two Dog worked. But he didn’t just sit back; he often argued with Two Dog over fragmented market information and AI-derived conclusions, "His attitude was terrible, and he thought I was just a young kid who didn’t know anything, so we couldn’t build trust. In the end, after some negotiations, we removed him."

"Honestly, our LPs are all very accomplished in their respective fields—they are authorities in their domains. But now, with AI as an assistant, they also believe they have authority in investing," Two Dog lamented.

Because fundraising channels are limited, LPs of small dollar private equity funds are mostly friends or acquaintances of the fund managers, making their composition quite complex. According to Two Dog, their fund’s LPs include high-net-worth individuals in China, business owners, and FOFs (fund of funds). "Our LPs include coal barons from Shanxi, wealthy individuals ranked in the top 300 on Forbes, and some LPs are even second-generation wealthy who were introduced by their parents."

Their relationships with LPs are also quite delicate. For some LPs, the fund might not even charge a 2% management fee, only taking 20% of profits. The main characteristic of this LP structure is their enthusiasm for participating in financial markets and "capital going abroad," but they lack the time and energy to quickly learn and research market trends.

In a sense, the core value of GPs is to handle information gathering, market research, opportunity screening, and investment judgment for LPs, compensating for their lack of time, energy, and cognition with professional expertise, thus transforming information into decisions.

However, with the proliferation of AI tools, this reliance on professional institutions’ information processing and research capabilities is rapidly being equalized. Aside from the final capital allocation and trade execution, much of the traditional work of GPs is beginning to be replaced by AI at lower cost and higher efficiency.

"Getting an IBKR brokerage account for our LPs isn’t hard. With AI assistance, they can now buy whatever industries or assets they like, entirely on their own," Two Dog believes that AI’s impact on funds using subjective strategies is especially significant because investment is always result-oriented. If LPs hit the right trend and their personal returns surpass the fund’s, they will naturally start questioning the fund’s strength.

In contrast, the "information equality" brought by AI has less impact on quantitative private equity funds and might even widen the gap between funds.

Parameters and algorithms in quant strategies are constantly iterated, and AI accelerates this process. This is a field where efficiency and intelligence compete. Without specialized knowledge in math or finance, ordinary people cannot build competitive quant strategies comparable to large quant funds using AI.

"Quant strategies fundamentally require constantly staying ahead of market peers to generate excess returns. If you think your ordinary AI-built strategy is good, it’s likely that most smart people have already discovered and iterated on it," Two Dog said, highlighting the advantage top-tier quant funds hold.

Will AI replace GPs?

But Two Dog is not worried that AI will truly replace GPs or analysts, because AI is always neutral and accessible to everyone. It’s a lever that GPs can use to improve their knowledge systems and investment strategies, creating more returns for LPs. What really frustrates him is that AI increases friction between GPs and LPs.

"Some LPs even question why you didn’t invest in the current hot assets and analyze everything thoroughly. They don’t understand that GPs don’t just invest in whatever is trending now," Two Dog said, somewhat speechless, especially after this year’s rally in U.S. stocks in AI and semiconductors, where retail investors’ bets on sector leaders could generate outsized gains.

In a bull market, retail investors’ returns can easily surpass those of funds. First, individual investors are more flexible, tolerant of mistakes, and more focused with their capital; second, with AI-assisted research, retail investors’ research efficiency is greatly improved—like having a versatile expert on call 24/7.

Especially in this year’s U.S. stock market, if retail investors hit the likes of SanDisk, Micron, SK Hynix, their returns could surpass most funds. "At this point, LPs might consider putting more into their own accounts and less into funds, or even withdraw from subjective private equity funds," Two Dog said. "In a bull market, everyone tends to think they are 'stock gods'."

But all this depends on retail investors using AI correctly. If they use subpar AI, the results will be half as effective. Two Dog said this is also the main reason for friction with LPs. "In China, high-net-worth individuals mainly use Doubao-like conversational AI, while more powerful analysis tools like ChatGPT or Claude are not yet widespread. This kind of companion AI, designed to provide emotional value, can easily create hallucinations in professional fields."

Essentially, the problem isn’t the AI’s capability level, but that most people don’t truly understand how to use AI. AI can integrate vast amounts of information in seconds and build a logically consistent analysis framework, but logical consistency doesn’t mean it aligns with facts. For LPs lacking professional backgrounds, it’s often hard to tell which conclusions are based on real data and which are just probabilistic inferences generated by models.

Therefore, most investors are seeking affirmation from AI rather than genuine analysis. The ultimate goal of AI isn’t to help investors "seek truth from falsehood," but to facilitate dialogue.

So, will AI replace GPs? AI can generate ten thousand logically coherent research reports at low cost, but asset management is fundamentally an "ancient service industry" based on trust and mental commitment. The process of mutual selection between GPs and LPs remains.

But in any future where "tasks" are ultimately executed by AI to maximize "results," even "human private equity" will need to learn from AI and further cultivate emotional value.