Huahuan "Technology Brother" counters "AI Bubble Theory": This time is different from the past, cash flow and engineer dividends are important supports

Hu Yibin, Co-Chief Equity Investment Officer of Huaan Fund and "high prosperity track layout builder," shared insights at the Huaan Fund "Year of the Horse Investment Carnival," focusing on leveraging the engineer dividend to invest in the future.

Overall, he believes that the A-share market is stable with upward potential and will still present structural opportunities, even a structural bull market.

In this context, he is optimistic about technology-related directions, including computing power, innovative drugs, independent innovation, and AI applications, all representing the engineer dividend. Additionally, there will still be concerns about PPI turning positive or other styles disrupting technology, leading to a search for configurations in cyclical and new consumption tracks.

Key Quotes:

1. As the engineer dividend is gradually released, R&D (Research and Development) expenditure as a proportion of GDP will also increase, which is a cumulative variable that will bring about China's future opportunity to occupy the global technological high ground.

2. If we make some horizontal comparisons with the averages of major economies and developed countries, our country currently has a significant advantage in the large number of university students related to STEM (Science, Technology, Engineering, Mathematics).

3. Currently, if we also experience a round of engineer dividends, more outstanding stocks will emerge in the fields of innovation, semiconductors, automobiles, consumer electronics, and precision manufacturing.

4. Compared to global leading companies in the same industry, the gross profit margin levels of A-shares differ significantly, possibly by two to three times, but this does not mean that such a gap is sustainable; both will regress towards the median.

5. Therefore, the profit margins of Chinese companies going overseas are underestimated. At this stage, revenue is increasing without profit because companies are currently expanding their territories, and in the future, there may be a Davis double-click type of explosion.

6. Today we saw excellent open-source products like DeepSeek and Alibaba's Tongyi; this is not a coincidence, as it is due to the large number of AI talents in China, and this talent density is closely related to the engineer dividend we mentioned.

7. The embodied intelligence industry is currently in the 0-1 stage, where the focus is more on product refinement and innovative supply. As this stage passes, we will see a steep growth rate, with demand potentially increasing tenfold each year, indicating a vast long-term space for this industry.

8. The per capita consumption of Chinese people may experience three pulse explosions, significantly increasing per capita consumption. The first is around the college entrance examination; the second is during career advancement, around the ages of 30-35; and the third is after retirement in the silver-haired era.

9. The differences in consumption concepts and products between our generation reaching 30, the previous generation at 30, and the next generation at 30 have formed the essence of new consumption.

Overseas "Engineer Dividend" Model

Having experienced the ups and downs of the market in 2025, which was stable yet rising, and having gone through these challenges, we now look forward to the overall investment strategy for 2026. Today, I still want to share my thoughts focusing on the engineer dividend direction, while also extending some investment opportunities and sharing my personal investment strategy for 2026 First, we need to review how the engineer dividend is generated.

Let's take Japan as a historical reference. Japan's GDP accounted for about 18% of the global total in 1995, reaching its peak, and then it continued to decline. However, the share of Japan's manufacturing value added in the global market did not peak in 1995.

In fact, the highest point of Japan's manufacturing value added share occurred between 2000 and 2005. The peak of Japan's manufacturing value added was around 2005 to 2010. This means that even in the nearly 10 years after Japan's GDP share peaked, Japan's manufacturing share continued to rise.

We believe that this decade was the result of the engineer dividend, which offset many other economic issues in Japan.

Behind the engineer dividend is the rapid growth of the total number of higher education graduates in Japan. Although we see that Japan's total labor population may have become very stable or even declined after 1995, it did not affect the supply of higher education or high-quality talent in Japan. After 1995, Japan still had a very large number of higher education graduates, leading to a very positive outcome. Japan continuously supplied high-quality talent to the labor population.

China has the most high-quality graduates in the world

If we make some horizontal comparisons with major economies and the average of developed countries globally, we can see that our country currently has the greatest advantage in STEM (Science, Technology, Engineering, Mathematics) related university graduates.

Compared to any other country or economy in the world, even when the 27 EU countries are combined, we have a higher total number of graduates. Of course, this includes vocational colleges and technical schools. Overall, our country sends about 12 million graduates to various industries each year, and in many research-related fields, we have a very ample supply of "fresh blood."

With so many graduates being "rolled over" each year, in many fields, such as high-tech areas, new energy vehicles, quantum computing, robotics, artificial intelligence, and so on, we have the largest number of high-quality graduates in the world, far exceeding the average of developed countries, which is about 3.2 million per year.

Currently, our density of engineers is relatively low compared to developed countries, with a slight gap, but we have a significant advantage compared to many other developing countries.

The gap with developed countries also means that the release of our engineer dividend may just be beginning. Our R&D (Research and Development) expenditure as a percentage of GDP is only slightly above the average of the 27 EU countries, but we are rapidly catching up with the United States, which is at 3.55, while we are at 2.68.

As the engineer dividend is gradually released, I believe that the proportion of R&D expenditure to GDP will continue to rise, which is a cumulative variable. As the engineer dividend continuously contributes new research breakthroughs to our country, such as DeepSeek we saw at the beginning of 2025, and the innovative drugs that followed, We see many new technologies going overseas, which are actually backed by a large number of STEM graduates entering the industry. With continuous accumulation, R&D expenditure as a percentage of GDP is increasing, transforming into advantages in scientific research results, thereby creating a context for occupying the global technological high ground and generating new investment opportunities in the future.

China Will Not Experience a "Lost 20 Years"

There are concerns about Japan's "lost 20 years," during which Japan experienced continuous price declines for about 20 years after 1990, and the economic stagnation was evident.

But history serves as a mirror. Even during those "20 years," we still saw the Japanese stock market continuously giving birth to "big bull stocks." For example, in the fields of innovative drugs, semiconductors, automotive consumer electronics, as well as precision manufacturing and chemicals, countless stocks have significantly outperformed the Nikkei 225 index.

These stocks are a reflection of the scientific research high ground brought about by the engineer dividend in Japan at that time, and ultimately occupying the global high ground. The performance of these companies was outstanding, detached from the domestic economic situation in Japan at that time.

We are likely to fare better than Japan, but from the perspective of discovering investment opportunities, the aforementioned history can inspire us. It tells us that seizing these opportunities can allow us to significantly outperform our benchmark, which is a question we are currently contemplating.

At present, if we also experience a round of engineer dividends, in fact, the foundation of our country's engineer dividend may be larger and better than Japan's. We may see more outstanding stocks emerging in the fields of innovation, semiconductors, automotive, consumer electronics, and precision manufacturing.

If we extend the timeline a bit further, from 1990 to 2025, during Japan's big bull market in the 15 years after 2010, as long as we provide a little "sunshine," the excess returns of these bull stocks will become more apparent, even surpassing index gains by dozens of times. This gives us a very good insight: as long as we can identify it as the best symbol of the engineer dividend in the medium to long term, the profits of these companies will significantly outperform the index during relatively unfavorable global macroeconomic conditions; however, when there are opportunities and the economy takes off globally, they will perform exceptionally well.

Leading Companies Going Overseas Will See an Increase in Gross Margin Levels

A common characteristic of these bull stocks is their relatively high gross margin levels. Why does this result occur?

For example, in the stock market cycle we observe, the company with the largest increase in Japan is Keyence, a company engaged in equipment and precision manufacturing. From 2010 to 2025, this company maintained a gross margin of over 70% and steadily climbed; its performance also steadily rose, which means that the company's moat is very high. Even with such a high gross margin, its scale is still expanding, and there are still no competitors that can impact its gross margin level, reflecting to some extent the engineer dividend in Japan, which has provided the company with a medium to long-term R&D advantage. This R&D advantage brings a very obvious moat What is the current state of China? If we compare the gross profit margins of several major industries with a high proportion of overseas operations, such as materials industry, communication services, information technology, and healthcare, with their counterparts in the US and Japan, we will find that, except for the materials industry, A-shares outperform their American peers in these overseas industries.

In all other major categories, the gross profit levels of A-shares are relatively low, especially in communication services, which are quite low, as well as in industrial sectors and pharmaceuticals, including innovative drugs.

This indicates that we are currently in the early stage of the engineer dividend explosion. A very typical characteristic of the early stage is a high degree of internal competition. Most of the industries going overseas are in a phase of gradually catching up in research and development, and some industries have already surpassed their counterparts, such as new energy battery technology, which has basically surpassed and formed a global high ground in research and development, but still needs to compete for global market share.

At this stage, we can see that most Chinese companies rely on price to gain market share after going overseas. Therefore, in this phase, land grabbing is more important than profit margins, so initially, our gross profit margin levels are indeed low. In the mid to late stages, it is possible for gross profit margins to rise, and I believe this is feasible.

Compared to major global leading companies, the gross profit margin levels of A-shares in the same industry vary greatly, possibly by two to three times, but this does not mean that such a gap is sustainable.

In the future, I believe there will be some significant investment opportunities, often referred to as the "Davis Double-Click," where the double-click refers to the growth in company scale and, when the scale increases, we will definitely see the company's profit margin levels return to the global leading companies' average.

Perhaps we currently see companies increasing revenue without increasing profits, but we need to be tolerant of profit margins at this stage. Therefore, we should theoretically give some leading companies a higher valuation because their profit margin levels are underestimated, simply because they are currently in a land-grabbing phase.

Looking back in 5 or 10 years, as long as the engineer dividend continues to be released, when the engineer dividend transforms into a global high ground in research and development and global technological monopoly, we will see that our companies, in addition to scale growth, will definitely experience a leap in profit margin levels, leading to performance that can consistently exceed expectations at certain stages.

The Chinese Engineer Dividend Supports Global AI Development

Let's return to a few directions for expansion.

First, from a technological perspective, this year, several major cloud giants in the US have poached some Chinese scientists with high salaries. On the globally renowned open-source platform GitHub, the contribution rate of Chinese and Chinese-descended individuals exceeds 50%. Based on the proportion of core technical personnel in the global Top 100 AI companies, the combined percentage of Chinese and Chinese-descended individuals is 42%. In terms of top AI papers, the proportion of Chinese and Chinese-descended individuals is close to 48%.

Undoubtedly, this year we have seen excellent open-source products like DeepSeek and Alibaba's Tongyi, but the emergence of such products and their occupation of the global open-source high ground is not a coincidence. It stems from the fact that China has a large number of AI talents, which is closely related to the engineer dividend we just mentioned Because we have 12 million STEM graduates every year, these graduates are constantly moving forward day and night, exploring various high points in the field of AI.

If there are new technological breakthroughs in AI, these 12 million graduates each year will work tirelessly to replicate and surpass them.

Therefore, from certain perspectives, our engineering dividend will definitely yield a result. In the end, the contributions of Chinese and Chinese descendants to global AI will continue to rise. There is no doubt that in a few years, the definition of AI or AI algorithms may largely be in the hands of Chinese or Chinese descendants, which is highly likely.

AI Will Have a Positive Cash Flow Cycle

We also see that throughout 2025, especially in the second half, AI's performance is outstanding. This performance mainly comes from overseas, with the emergence of generative AI.

So far, we have seen for the first time cloud payment revenue or annual subscription revenue, which forms a positive cycle with the enterprises' computing power investments.

When AI companies are making capital expenditures, purchasing computing power cards, and investing in large-scale computing energy infrastructure, they can simultaneously convert computing power into profits and payments for enterprises. This is the first time since the emergence of generative AI in 2023 that we have seen an explosion in annual subscription revenue.

At the same time, we will see these companies' cash flow steadily increasing, with some companies continuously raising their expectations. This has led to a situation where, whether in inference, ASIC, or training cards, the demand currently appears to be in a state of rapid growth.

We have also seen recent market concerns about a bubble in AI, often comparing it to the dot-com bubble of 2000.

I just want to make one point to refute this. In the dot-com bubble of 2000, we saw capital expenditures exceed net cash flow from operating activities. The investments in the U.S. stock market at that time were not based on the rapid growth of operating cash flow.

In other words, at a certain point, capital expenditures exceeded cash flow, leading to rising liabilities. Therefore, when liquidity declined, the company's profitability could not support the continuation of capital expenditures, resulting in a rapid peak of the dot-com bubble.

However, what we can see now is that in the AI market, the capital expenditures and cash flow of Nasdaq companies are showing a trend of simultaneous increase, and we can even see that in the short term, due to the AI explosion, operating cash flow far exceeds the growth of capital expenditures, so there is currently no need to worry about the companies' debt issues.

Differences Between Domestic and Overseas

Overseas is like this, but what about domestically? Domestically, we can see leading companies, such as ByteDance, where AI products have not significantly generated revenue. Top-ranking products like Doubao are currently almost free, but this does not mean they lack commercial value China's business model is quite different from that of the United States. China has traditionally focused on value-added services or indirect business models, having invented many indirect value-added service business models in the internet space. In fact, China's foundation for internet payments is not poor, with online games, live streaming, short dramas, online literature, and so on.

Globally, the origins of business models can be traced back to China, so China's consumer-facing internet payment business foundation is not weak. However, we did not simply and bluntly charge users directly through subscriptions in the early stages; instead, we focused more on cultivating users' long-term usage habits and then obtaining revenue through indirect or direct payments when user stickiness is relatively strong.

Is there no commercial value at present? Not at all. During the development of the internet, we see that the valuation of internet companies mainly comes from active user numbers, retention, and their ARPU (Average Revenue Per User), etc.

The best way to view internet companies at present may be through the growth of tokens. The consumption of tokens directly represents users' stickiness to the AI products of internet companies. The growth of tokens we can observe now almost doubles every 4 to 5 months.

Therefore, based on this foundation, we firmly believe that the future value of China's internet companies, including these CSP companies (Cloud Service Providers), remains promising in the medium to long term.

The Endpoint-Cloud Hybrid Direction is Worth Attention

I would like to mention a few other directions. One is the endpoint-cloud hybrid. We believe that AI ultimately needs to solve the computing power issues on the endpoint side while also addressing user experience. We need high-emotional-intelligence large models, but these large models cannot be placed locally because they are very large, and local computing power cannot support them. At the same time, we need local solutions because there are many personal privacy data points locally.

I think AI models should undergo some form of embedded model training locally or import personalized data locally, but these personalized data points are often difficult to upload to the cloud. Therefore, I personally believe that the future trend is endpoint-cloud hybridization and the integration of cloud and edge, forming an AI model.

Other Directions Reflecting Engineer Dividends

Apart from AI, there are many other manifestations of engineer dividends in the technology sector.

For example, in the embodied intelligence industry, we are currently in the 0-1 stage of the humanoid robot industry, where the focus is more on product refinement and innovative supply. As long as there is supply, there will be demand. Therefore, during these years of significant investment in product development, China is actually number one globally in terms of both the number of patents and the number of supporting enterprises. As we move past the 0-1 stage of the industry, in the 1-10 stage, we will see a steep growth rate where demand may increase tenfold each year because the long-term potential of this industry is very large. In this stage, we will witness a Davis double-click on PE and EPS New Energy, without a doubt, we have formed a globally leading advantage in solid-state batteries, including photovoltaic perovskites, with fast technological iteration, early progress, and strong industrial synergy. In the future, as the application fields of solid-state batteries expand from automobiles to other areas, our leading enterprises will have an even better competitive landscape globally.

In the field of innovative drugs, we have a rapidly growing drug pipeline that holds a first-in-class position (the best in the world). Our pharmaceutical R&D talent is doubling every five years, so we are currently the number one in innovative R&D talent globally. Our patents are nearly on par with the United States, and perhaps in the next 5 to 10 years, we will see a significant number of globally innovative drug products or major products originating from Chinese enterprises.

The New Consumption Source is Generational Differences

We do not believe that there are no opportunities in other market directions. In areas not closely related to the engineering dividend technology industry, we are also seeking opportunities to smooth out the excess volatility of our portfolio.

I would like to mention a few points in other directions. The first is new consumption. Over the past 20 or 30 years in Japan, we have seen that besides finding investment opportunities with an engineering dividend explosion, there have also been some very good investment opportunities arising from affordable consumption and new consumption trends.

Chinese per capita consumption may experience three pulse explosions, significantly increasing per capita consumption. The first is around the time of the college entrance examination; the second is during career take-off, around the ages of 30 to 35; and the third is after retirement in the silver-haired era. The highest income stage in our lives may be between the ages of 30 to 45 or up to 50, but during this stage, there is not much time to consume, so the consumption tendency is not the highest at this stage.

Our consumption capacity is a downward U-shaped curve, while the consumption tendency is an upward U-shaped curve. When these two are combined, if we take the difference, we find that it results in three consumption pulses in life. When we encounter these three consumption pulses, it creates the most important investment opportunities.

The differences in consumption concepts and consumption products between our generation reaching 30 years old, the previous generation at 30, and the next generation at 30 form the essence of new consumption.

In the entire medium to long term, when we invest in the new consumption context, the most important thing is to grasp the differences in consumption concepts and consumption products brought about by generational iteration during the three consumption pulse stages in life, which is our investment opportunity.

Anti-Involution and Technology

Another direction I want to mention is anti-involution. In many medium to long-term industrial products, the overall gross profit margin and net profit margin levels of our manufacturing industry are relatively low. Anti-involution effectively breaks through the low profit levels caused by extreme involution in domestic enterprises.

I believe that the impact of anti-involution on the national economic price system is quite profound. In the medium to long term, it may represent a turning point for corporate profitability and the downward trend in employment and product prices faced by enterprises We can also see that our current PPI has experienced a continuous decline for 4 years. Since the reform and opening up, the relatively long PPI down cycle has been about 4 years, so there is reason to believe that as we move forward, the possibility of PPI returning to positive territory is increasing.

We can see that the prices of goods related to global pricing have already shown very prominent trends. Domestically, we are undergoing economic structural transformation, so most products related to real estate are still in a downward range, but this does not mean they will decline permanently.

Finally, I would like to mention a direction, which is the Hang Seng TECH Index. Compared to the Wind All A Index, its excess returns have been retraced for some time, and the index level has returned to the beginning of the year. It has now returned to a historically very low level, which also represents the pricing differences between domestic and foreign capital for Chinese assets. This difference will be corrected in the medium to long term.

Assuming that next year we really see PPI and various asset prices turning positive, and China successfully emerges from price declines due to economic structural transformation, I believe the Hang Seng TECH Index will also return at that time.

Risk Warning and Disclaimer

The market has risks, and investment requires caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial conditions, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investment based on this is at one's own risk