Jinglin Asset's year-end statement: 2026 will be another pivotal year for AI, with international funds returning to Chinese assets.

```

On December 29, Chinese private equity giant Jinglin Asset Management sent out its “year-end letter” to investors.

Zishitang has learned: Jinglin Asset fund manager Gao Yuncheng stated in the letter, “2026 is very likely to be the inaugural year when AI Agents truly become widespread.”

Gao Yuncheng pointed out in the letter: AI large models that use hardware such as mobile phones as the underlying operating platform may even challenge existing iOS and Android systems. Many of the established moats may be overturned in this round of technological change, and companies without an AI ticket will be marginalized. This is something we need to be alert to!

Zishitang summarizes the main points of the letter as follows, for the readers’ benefit.

Core viewpoints:

1. The fact that 50% of AI researchers are Chinese is actually a recognition of China’s basic education. China has both competitiveness and cost-performance advantages in the generative artificial intelligence competition wave.

2. In such a macro environment, we observe that international capital is returning to Chinese assets. This process has just begun.

3. As investors, we should not casually draw conclusions about AI, but should continuously scrutinize the cash flow status of companies in our portfolio, and recalculate valuations based on stricter accounting assumptions, as this is what we need to do.

4. The valuations of our core portfolio companies are all within historical reasonable ranges, their business models possess pricing power, the competitive landscape is favorable, and management is sufficiently clear-headed and rational.

5. Our core portfolio is basically stable, and differs little from the businesses shared with everyone in the middle of the year. In a nutshell: two systems, each with its stars.

6. Only companies with outstanding business models and especially unique competitive positions can control profitability. Such companies are extremely rare.

China Has Competitiveness in Generative AI

The debut of DeepSeek in 2025 allowed the world to recognize China’s competitiveness and cost-performance advantages in the generative AI competition wave.

It’s not just DeepSeek—Qwen’s global adoption among open-source models also proves that even in the absence of the most advanced GPUs, domestic Chinese models can utilize available resources to make leading competitors anxious, and give the world an option beyond the US.

The fact that 50% of AI researchers are Chinese is actually a recognition of China’s basic education system; the large number of STEM talents truly supports continuous upgrades from basic industry to aerospace and military technology.

The generational iteration of engineers behind the development of China’s industries fully explains our historic process: from learning from western advanced manufacturing, to participating in complex global supply chains, to surpassing others in R&D and manufacturing in some industries.

With a strong talent pool and already-developed globally competitive industry leaders, China now has sufficient strength.

International Capital is Returning to Chinese Assets

In this macro context, we observe that international capital is returning to Chinese assets. This process has only just begun.

In the frontier fields of new economy such as AI, new energy, intelligent driving, and humanoid robots, both China and the US are investing actively, and related companies are growing rapidly in this wave, with their revenues and valuations constantly setting new records—the future is as vast as the stars and sea.

Be Cautious With Overly Optimistic Expectations

Yet, as this rapid development occurs, dissenting voices are increasingly heard.

Discussions about whether future cash flows can support trillions in AI-related infrastructure investment kept surfacing at the end of 2025. The capital expenditures—on and off balance sheet—of major tech companies have grown so large that many question whether the cycle can be sustained at some point.

As investors, we shouldn’t draw conclusions lightly. But in light of such risk warnings, we need to continually scrutinize the cash flow status of companies in our portfolio and recalculate valuation levels based on stricter accounting assumptions.

Indeed, some companies’ market values are built on extremely optimistic expectations for the future, but these are not the companies in our portfolio.

The valuations of our core holdings are all in historically reasonable ranges, their business models have pricing power, they are in a favorable competitive position, and management is sufficiently sober and rational.

Future Giants = AI-Native Companies Just Starting Out

The regret this year is that we didn’t actively engage in the storage and information transmission segments of the AI industrial chain. The reasons: the former was overly focused on cyclicality and overlooked structural demand changes and the difficulty of adding new capacity; the latter was due to worries about excessively high profit margins and changes in future technology paths.

While mindful of bubbles or overinvestment, we can also clearly see that AI’s penetration and transformation of all industries has only just begun. The giants of the future may still be AI-native companies that are just getting started. Once such highly efficient tools are brought into use, there’s no going back.

The main differences: which AI tools to use? How much to spend? The research efficiency and cost-performance shown by China’s leading companies have once again proven themselves to the world.

Outside these fields, the global economy actually faces pressure for growth as well, including employment pressures. How to maintain low unemployment in the age of AI is a global challenge.

There are hopes, there are problems, there is feverish concentration of resources, and there are deep freezes that will take time to thaw—these layers of complexity, plus human greed and fear, manifest in the stock market as high individual stock volatility—holding firm is not easy.

Many investors sigh that if only they had simply held the companies they owned at the beginning of the year until year-end, returns would have been far higher than through active trading. That’s the difficulty of investing.

Two Systems, Each with Its Stars

Our core holdings are basically stable, with little difference from the businesses in the portfolios we shared in mid-year.

In a nutshell: two systems, each with its stars.

In the two major systems dominated by the G2, we select companies with reassuring valuations and operating cash flows according to each system’s comparative advantage.

All companies in our core holdings share this common feature: they have strong client stickiness and pricing power, obvious product differentiation, and can control profit margins at their own pace.

As I mentioned in a previous letter, the vast majority of companies are running desperately in the competitive market, with no composure in controlling the tempo.

Only companies with outstanding business models and especially unique competition patterns can control their profitability. Such companies are extremely rare, and we may continue to hold them until they start to lose this privilege. This is also a key feature of companies with real business moats.

In addition, we should pay close attention to companies that are important entry points or platforms for AI applications.

2026 is Very Likely to Be the Inaugural Year of True AI Agent Proliferation

Companies that can be called world-class entry points or platforms number only a few, such as Google, Meta, Apple, ByteDance, Tencent, OpenAI, etc. 2026 is very likely to be the inaugural year when AI Agents become truly widespread.

Software-based AI Agents and integrated AI Agent devices will continue to be launched. AI large models that use hardware such as mobile phones as their underlying operation platform even have the potential to challenge the existing iOS and Android systems. Many existing moats may be overturned in this technological wave, and companies without an AI ticket will be marginalized. This is something we need to be wary of!

“Evolution” Within Private Equity Funds

After developing for more than 20 years, Jinglin can now report to our investors: we have now preliminarily established the capability for research and investment in some key overseas markets.

On the basis of being rooted in China background-listed companies, we’ve spent 7-8 years building teams and systems, and in some industries now gradually have the ability to find the best companies in the world along the industry chain.

Risk Reminder and DisclaimerThe market has risks; investment needs caution. This article does not constitute personal investment advice, nor does it take into account any individual user’s specific investment objectives, financial situation or needs. Users should consider whether any opinions, views, or conclusions in this article fit their particular circumstances. You are solely responsible for investment decisions based on this article. ```