Anfu, Dongshan, and Fosun have imitated one after another. Why can’t Berkshire be created in China?

This article may contain an automated translation. The original source is linked for verification.

Anfu, Dongshan, and Fosun have imitated one after another. Why can’t Berkshire be created in China?

brocade

2026.11.11 11:38

· From Beijing

Full text 6765 words

Text | brocade

In October, global AI assets collectively pulled back. The Philadelphia Semiconductor Index fell by more than 15% in a single month, and the Nasdaq 100 Index fell by nearly 7%.

On the other side, there is the representative of the "old era" - Berkshire Hathaway, whose stock price hit a record high on October 22. In the past month, its Class A shares have risen by more than 12%, while the S&P 500 has fallen by 5.5% during the same period; it has risen by 13% year-to-date, also outperforming the S&P 500 by 7.8%.

Most explanations of this combination on the Chinese Internet are: value investing is back, and Buffett wins again.

In fact, this is not true. Berkshire is no longer a bystander of AI. After Buffett stepped down as chairman, Google rose to its top three holdings and invested more than 10 billion U.S. dollars in AI-related assets in the third quarter. However, Google has fallen by more than 24% since June and has already fallen below its cost.

This shows that Berkshire’s new high is not a victory of “avoiding AI”, but a victory of the mechanism itself. So, the old question resurfaced again: Why doesn’t China have Berkshire Hathaway?

In fact, there are many companies in China that want to learn from Berkshire. Guo Guangchang has publicly said that he wants to be China's Berkshire; Youngor has been a troika of "clothing, real estate, and investment" for more than 20 years; industrial capital representative Yuan Yonggang used a series of mergers and acquisitions to push Dongshan Precision to a market value of more than 500 billion yuan, and also established Anfu Technology.

Counterintuitively, China does not lack "Buffetts who can dig up good companies." In a highly volatile market like A-shares, there are still many investors who can make long-term returns of more than 15% annually, and their stock picking abilities are no weaker than Buffett's.

Therefore, the real crux of why many cases are not successful is: Berkshire, a sophisticated machine, requires three conditions on the liability side, asset side, and governance side to be established at the same time; and in China, no company can build the same structure, and it will be difficult to estimate the future.

Let’s figure out Berkshire first

If you think of Berkshire as a "company" or an investment portfolio, you will never understand it. In fact, it consists of three businesses: more than 50 industrial companies and more than 40 stock portfolios + property insurance business of 14 insurance brands + cash.

A more accurate analogy is that Berkshire is a very special perpetual closed-end fund, cleverly designed in three aspects: the liability side, the asset side, and the manager.

1. Liability side: negative cost, and basically sustainable

The market often says that "Buffett uses insurance and leverage to speculate in stocks." This understanding is wrong. Berkshire's total assets/net assets are about 1.6 times, and the leverage of the insurance sector is about 1.8 times, which is lower than the average of 2.4 times in the US property and casualty insurance industry. Its leverage is not high, but the quality of its liabilities is high.

The insurance business contributes about one-third of Berkshire's profits and is the core source of the company's debt advantage. The float cost of its subsidiary GEICO (one of the largest auto insurance companies in the United States) in the past 20 years has been -7.8%, corresponding to a 20-year average combined ratio of 92.2%. The underwriting is profitable in the long term, and the policy liabilities can be paid off in the short term - you invest other people's money, and others pay management fees.

Coupled with the holding company's operating cash flow, investment assets and U.S. debt interest, this part of the cost of funds is also negative.

2. Asset side: a business that can collect rent, and try to buy it as a whole

What Berkshire buys is not stocks that will rise, but assets that can collect rents. There are roughly three categories:

Regulated monopolies - BNSF Railway (North American first-class railway, non-replicable right of way, controlling nearly 30% of freight volume on the West Coast of the United States), BHE Energy (power grid and power generation, returns subject to regulatory approval, covering power supply in 16 states in the United States);

Brands with pricing power - Apple, See's Candy, Coca-Cola, American Express, Nippon Trading Company;

Services of local monopolies – Clayton Homes, Benjamin Moore, NFM, American Express.

These have in common: demand is rigid, prices are determined by regulation or branding, cash flow is predictable, and we try to buy them as a whole. As for whether these assets have high growth, it is not that important. What is more important is that they can have reasonable prices and have long-term sustainable operations.

For example, UBS evaluates BHE as "a largely regulated investment base" and gives an assumption of only 4% net profit growth in the next five years. It buys duration and cash flow - using zero cost money to buy a lease that lasts for 50 years and has a slight price increase every year.

Moreover, Berkshire’s purchase of BNSF is 100% equity. In order to control and hold it for a long time, it needs decision-making rights, not just dividend rights.

Another case that is less discussed by the market but I think is very interesting is that Berkshire’s AI assets include more than just Google. As early as January 2016, Berkshire spent US$37.2 billion in cash to wholly acquire PCC, one of the two leading companies in gas turbine blades, and became its subsidiary and delisted.

This was the largest acquisition in Berkshire's history, and later a goodwill impairment of approximately US$11 billion was accrued due to the downturn in the aviation industry. However, with the explosion of gas turbines for AI power generation and the recovery of aviation demand, the public market value of another leading company, HWM, has exceeded US$80 billion, which means that PCC may also have the same market value.

Even though it was a "major mistake" that Buffett publicly admitted, he actually achieved more than double the income in 10 years; Buffett's negative cost advantage and redundant cash on the liability side enable it to withstand the drawdown. More importantly, as a monopoly asset, it is only a matter of time before the capital is recovered.

Therefore, most of the asset classes deployed by Berkshire can indeed travel through cycles. It is precisely to match the negative cost liability structure that Berkshire always values stability when buying assets. The purpose is to collect rents, and the purchase price is just to make the rental yield higher.

Figure: Portfolio of listed companies held by Berkshire Source: Announcement

3. Manager: authorization across terms

Managers are the most discussed, and Buffett and Munger are the most famous investors in the world. Let’s expand a little and talk about something that people don’t pay much attention to at the architectural level – cross-term delegation.

Berkshire's more than 50 businesses are managed by 28 individuals from its Omaha headquarters. The subsidiaries operate independently, and the group only allocates capital. The premise of this structure is that the person who allocates capital has multi-term authorization (Buffett has held more than 30% of the voting rights for a long time, and the equity ratio is about 14%), and is not responsible for single-year profits - the proportion of a single asset can exceed the upper limit of any portfolio management.

Therefore, this manager structure further amplified Berkshire's asset and liability advantages, and finally achieved the miracle of compound interest of 61,000 times.

Dismantling the reasons why it is difficult for domestic benchmarking companies to do so

1. The cost of debt has stuck most people.

Most domestic holding companies rely on bank loans, and the cost of capital is interest; in the insurance industry, due to extremely fierce competition in insurance policies, the capital cost of the entire domestic life insurance industry is above 3%.

Therefore, high-cost liabilities will deform the investment side's actions, and assets that can purchase high capital gains in the short term are the primary pursuit. This kind of institution with high cost ratio and high leverage is an investment myth in the upswing period, but it is easy to learn lessons in blood and tears in the downturn period.

If capital costs are high but you insist on continuous mergers and acquisitions, what dangers will you easily step into? HNA gave the answer. From 2015 to 2017, HNA completed more than 80 mergers and acquisitions totaling more than US$40 billion in three years, covering hotels, logistics, and banks. At the end of 2017, its total liabilities were 123.4 billion yuan. In 2018, it sold nearly 300 billion yuan of assets and liquidated more than 300 companies. In January 2021, it entered bankruptcy and reorganization, with debts of approximately 60 billion yuan in the aviation sector.

Even Fosun Group, which also has property and casualty insurance business, has a comprehensive interest-bearing debt cost of about +5.0%, which is 12.8 percentage points different from Berkshire’s float cost of -7.8%. Fosun’s profit expenses are approximately 27.9 billion yuan a year, which is 16 times more than the net profit attributable to the parent company of 1.7 billion yuan in the first half of 2026.

There is no cheap money, and mergers and acquisitions are not compound interest machines, but meat grinders.

2. Assets that can be rented cannot be bought or rented at a price.

On the asset side, China has quite a few assets that can be used to “collect rent”, some of which are even of better quality. Yangtze Power's 40% net profit margin, Moutai's 91% gross profit margin... Many companies are regarded as having a "money printing machine" business model.

But first of all, controlling stakes in such assets are not for sale. China Three Gorges Corporation and persons acting in concert hold over 50% of the shares, and Moutai Group holds approximately 54%. These ratios are not financial holdings, but the state’s control over energy, resources, and brands.

In the secondary market, you can only serve as a minority shareholder. You cannot join the board of directors, you cannot replace the management, and you cannot have the right to decide "how to use the rent." (Hillhouse tried to control Gree, but ended up with indigestion)

In addition, due to the need to protect people's livelihood in China, the rental yield of this type of monopoly business is actually capped. Changdian's on-grid electricity price is determined by the price authority, Shenhua's Changxie coal price is locked by policy, domestic mobile phone package rates are the lowest in the world, and highway tolls are strictly controlled: the rent collection rate is suppressed near the average social return.

It's also called "regulated", with the United States as a "bottom limit" and China as a "ceiling limit." These are two completely different rent collection structures. When Berkshire buys BHE, regulation gives a predictable rate of return; when you buy Changdian Electric, regulation gives you an electricity price that cannot be increased arbitrarily. The former is a lease and the latter is a utility.

Ultimately there is the difference brought about by valuation. Buffett repeatedly emphasizes that it is cheap because he has a huge amount of money and can wait for opportunities indefinitely on the debt side. Domestic investment institutions and holding companies are assessed on an annual basis and are not allowed to do nothing for many years. Therefore, the more it is recognized as a safe rent-collecting asset, the more funds push it to a high valuation premium.

3. Manager authorization boundaries are different

Even if it is negative cost float, China can find similar companies. For example, in the property and casualty insurance industry, the combined ratio of China Property and Casualty Insurance in 2025 is 97.6%. Calculated based on COR minus 100, the float cost is -2.4%, which is also a "negative cost fund."

Although this money can also be invested in stable rent-collecting assets, in addition to the above-mentioned prohibition of holding shares and only equity participation, the bigger problem is that the manager's authority as a capital allocator is limited.

Due to solvency and regulatory requirements, China Property & Casualty Insurance Co., Ltd. can only allocate limited assets in stock assets, and the latest figure is around 16%. This figure is actually close to the 19% average of the US property and casualty insurance industry, but it is significantly lower than Berkshire's level of more than 50%.

China Property & Casualty Insurance Company is an excellent property and casualty insurance company, but it is not and is not required to be an investment holding company. Accounting further amplifies this constraint. From 2023, listed insurance companies will implement new financial instrument standards. A large number of equity assets will be included in FVTPL. Every fluctuation in the fair value of stocks will be directly entered into the income statement, and the income statement will be assessed continuously. Insurance institutions are therefore institutionally more inclined to hold low-volatility assets rather than concentrate holdings like Berkshire.

Domestic case analysis can only resemble a certain aspect of Berkshire

1. Fosun International: The closest imitator, lost due to debt cost

Fosun is the only company in China that openly follows Berkshire's logic: insurance provides funds, investment provides returns, and the main focus is on the liability side, but the endowment gap is large.

Citigroup wrote very directly in the report: "When making acquisitions, Fosun pools its own funds with premiums from its insurance companies... premiums are long-term and provide a low cost of capital." Using premiums and its own funds to make acquisitions, the premiums have a long term and low cost. This is the Fosun version of the Berkshire model.

The quality of its insurance business is actually pretty good: in the first half of 2026, the insurance segment revenue was 22.807 billion yuan, and the profit attributable to the parent company was 1.245 billion yuan; Fosun Portugal Insurance has a market share of 30.1% in Portugal, ranking first; the underwriting end's combined ratio is about 90%, which is profitable.

But the scale is not enough. Berkshire's float is US$171.6 billion, corresponding to total assets of US$1.2 trillion; Fosun's total premiums for the first half of the year were 53 billion yuan, but the group's total debt was 218 billion yuan. Premiums can only cover a small part of the acquisition capital, and the rest must come from debt.

So the whole story slipped to the opposite side:

The gap is filled with interest-bearing debt. At the peak of interest-bearing debt in 2020, financing costs were 5.0%, a difference of 12.8 percentage points from Berkshire’s −7.8%;

The asset duration is long and the liability duration is short. As soon as the financing environment tightens, assets must be sold: in 2022, more than 40 billion yuan will be withdrawn from contract signings and 29.3 billion yuan will be repatriated; in 2023, nearly 20 billion yuan will be signed and 27.3 billion yuan will be repatriated;

The discount is self-reinforcing: PB is only 0.35 times, and the NAV discount is estimated to be 74.2%. The deeper the discount, the less likely it is to issue shares; the less likely it is to issue shares, the more dependent it is on debt; the more dependent it is on debt, the deeper the discount.

Fosun's current task is to "deleverage" and is far from the "compound interest" stage. It just proves that without the premise of negative cost funds, the same "insurance + investment" structure is likely to get the opposite result.

2. The other way is industrial capital, two samples of the Yuan family: Anfu Technology and Dongshan Precision

What industrial capital studies is the asset side—leveraging debt for mergers and acquisitions. On this road, Anfu Technology and Dongshan Precision, both controlled by Yuan Yonggang's family, have successfully "snake swallowed the elephant" in both cases, and the market attention is extremely high.

Figure: The stock prices of Dongshan Precision and Anfu Technology have fluctuated greatly in the past year.

(1) Anfu Technology: Use debt to buy cash flow, and then use cash flow to buy another

Anfu Technology does not learn from Berkshire's liability side, but its asset side logic - use debt to buy an asset with strong cash flow, and then use the cash flow from this asset to repay the debt and support the next investment.

In 2022, the company obtained the equity of Yajin Technology through two cash acquisitions. The company recently penetrated 46% of its holdings in Nanfu Battery, with a total transaction consideration of 3.75 billion yuan. The source of funds is the introduction of minority shareholders, bank M&A loans and shareholder borrowings.

Although Nanfu Battery has not grown for many years, it is actually an asset with quasi-monopoly attributes: the sales share of alkaline AA-5/AA-7 batteries in the retail market has exceeded 80% all year round, ranking first in domestic sales for 30 consecutive years, and the business is mature and cash flow is very good. It is a typical small and beautiful cash cow, very similar to See's Candy.

Figure: Anfu Technology’s history of holding Nanfu Battery Source: Company announcement, SDIC Securities

This style of play is actually more like the approach of Berkshire in its early days (the era of textile mills). But the hidden danger lies in the reinvestment after the cash flow is obtained. Debt pressure makes its investment one-sided to buy growth.

Anfu's reinvestment directions are: Yilanwei (holding 21% of the shares and the second largest shareholder) in thin-film lithium niobate chips, 10% of the shares of Huashengmai, a lithography machine light source company, and 6.9% of the domestic GPU Xiangdixian.

These are all highly volatile semiconductor growth assets that compete for growth arbitrage rather than rent collection. Naturally, the market will not give a premium for capital allocation. This debt structure also determines the risk of additional issuances at any time in Anfu Technology. The company's error tolerance space in investment is actually much smaller than that of Berkshire.

(2) Dongshan Precision: Treat "acquisition-integration-reacquisition" as a growth flywheel

If you look at the Yuan family's operations on Dongshan Precision, it will be easier to understand Anfu Technology. Dongshan Precision has gone even further, relying on stock issuance and bond issuance, continuous mergers and acquisitions, and constantly changing tracks, making mergers and acquisitions its main business, with a market value of more than 500 billion yuan at its peak.

When it went public in 2010, Dongshan Precision’s main business was precision sheet metal. In 2016, it spent US$611 million in cash to acquire MFLEX, the leading FPC company listed on Nasdaq, and entered Apple's supply chain in one fell swoop; in 2018, it acquired Multek to strengthen the rigid board; in 2025, it acquired Source Photonics for no more than 5.935 billion yuan, entering into 800G/1.6T optical modules and EML optical chips - every step of the way is on the cusp of technological iteration.

Objectively speaking, in terms of the investment value-added multiple on the asset side, Dongshan’s annualized rate of return is significantly higher than that of Berkshire: But the difference between it and Berkshire is precisely essential:

First, what it buys is not "rent" but "option". FPC, PCB, and optical modules are all industries in which technology routes are rapidly iterating—today’s share may be zeroed out by the new route in three years. Therefore, Dongshan’s mergers and acquisitions cannot stop and must keep buying a ticket. If Berkshire buys BNSF, it can stay put for 50 years. If Dongshan buys Source Photonics, it is betting on the 800G/1.6T wave.

Second, it has no negative cost sources of funding. Every acquisition involves real cash or financing, and the room for error depends on the attitude of the capital market. In a bull market, mergers and acquisitions are market value amplifiers; in a bear market, goodwill and debt are double-edged swords.

Third, its "compound interest" relies on management never making mistakes. If Berkshire buys the wrong company, the mechanism itself can absorb it; in Dongshan's model, a major misjudgment may interrupt the entire flywheel.

Therefore, looking at the two samples of the Yuan family together, it just spells out the whole picture of China's industrial capital: Anfu proved that "buying cash flow" works, but it can only grow due to debt costs; Dongshan proved that "buying growth" can make it bigger, but it cannot stop.

Conclusion

The essence of the Berkshire machine is the simultaneous establishment of three rare conditions: negative cost and sustainable money, rent-collecting assets that can be purchased as a whole, and multi-term capital allocation authorization.

The reality in China is: money is either expensive or restricted by authorization; assets are either not sold or the rent collection rate is capped; managers are tied up by annual assessments, dividend requirements and new accounting standards.

Therefore, "Why does China not have Berkshire Hathaway?" is not a question about capabilities, but a question about institutional conditions. So this also means that Chinese investment companies cannot become Berkshire, but can only be like a part of it.

For example, financial holding platforms, such as CNPC Capital, SDIC Capital, Minmetals Capital, COFCO Capital, etc., address the source of funds, but are restricted by capital adequacy ratios and consolidated supervision, and cannot achieve long-term cross-industry holdings.

Industrial capital platforms, such as Dongshan Precision, Anfu Technology, Youngor, and Nanjing Hi-tech, solve the problem of "investment" and cash flow reinvestment, but it is difficult to hold them permanently. Exit depends on IPO or dividends, and investment profits are not priced separately.

Insurance asset management companies such as Ping An of China and China Property & Casualty Insurance Co., Ltd. have negative cost float and long-term funds. What they lack is the authorization to “hand over the float to a single allocator.”

This article is based on public information and is for information exchange only and does not constitute any investment advice.

Respect for the original work, if you are qualified, you should be appreciated.

According to the real-name requirement of the Cybersecurity Law, please bind your mobile phone number before posting a comment.

Express