The "March Transition" has begun

I. From Expectation to Reality in March.

This week (20250224-20250228), the market structure not only showed a "high-low switch" in terms of price fluctuations and valuations but also exhibited characteristics of switching from the AI-led technology sector to traditional domestic demand areas. The reasons behind this have been detailed in our weekly report from the previous month, mainly three points: First, there has been extreme differentiation in trading heat and volatility among various sectors in the A-share market, which is expected to converge; second, the performance of technology stocks relies on the continuous emergence of new catalysts. Even when new catalysts appear, based on last year's fourth-quarter experience, the market's reaction to the catalysts themselves may gradually dull; third, the internal momentum of economic recovery is gradually becoming evident, while the logic of the artificial intelligence + robotics industry is facing marginal challenges.

II. After the "American Exception" Story is Broken.

The macro narrative of AI driving U.S. economic growth is facing challenges. The preliminary value of the U.S. February S&P services PMI recorded 49.7, hitting a 25-month low. The number of initial jobless claims for the week of February 22 exceeded the levels of the same period in the past three years. As an economy dominated by the service sector, the cooling of the service industry may indicate that the U.S. economy has not escaped its original cycle due to AI. Additionally, personal consumption expenditures in January fell month-on-month for the first time in 22 months, and the index of second-hand home contract signings dropped to a historical low. This at least indicates that a significant portion of the U.S. economy is still constrained by high interest rates. Against the backdrop of weakening economic data and Trump's continuous wielding of the "tariff stick," the Atlanta Fed predicts that the U.S. economy will shrink by 1.5% in Q1 2025, marking the first negative growth since Q1 2022. When the economy weakens, even the capital expenditures of major corporations cannot remain unaffected. Currently, the capital expenditures of the seven major U.S. tech companies account for nearly the highest proportion of their operating cash flow net, and recession expectations may drag down capital expenditures.

III. Both Overseas Economy and Inflation Face Headwinds, and Chinese Tech Stocks Also Experience Differentiation.

As the domestic economy in the U.S. weakens while inflation levels remain relatively controllable, the Trump administration has further elevated the priority of the "tariff card." However, its recent efforts to reduce inflation on the supply side have made little progress. The current inflation in the U.S. seems to be more a result of declining demand rather than supply-side improvements, and the uncertainty on the supply side has increased due to tariffs, ultimately pointing to either a weakening of demand or a rebound in inflation. Outside of the U.S., both Europe and the Bank of Japan have expressed caution regarding inflation, and a marginal tightening of overseas liquidity seems to be a more likely direction. Can Chinese tech stocks "stand alone" and become a safe haven for investment funds in technology? On one hand, from the perspective of overseas reflection, after this earnings release, Nvidia's valuation has been digested to 41.8 times PE, highlighting the "anchoring effect." On the other hand, based on historical experiences of typical industry cycles, A-share investors' rush to capitalize on industry cycles often does not exceed one year. In contrast, the performance elasticity of the rush combinations over the past two years seems to be "long overdue." Looking ahead, the performance realization of Chinese tech stocks may lead to market differentiation. **

4. Chinese assets are not only low-position defenses; the diffusion market has just begun.

Since the beginning of this year, the performance of the Hang Seng Technology Index has significantly outperformed the CSI 300 (which has historically been correlated with the relative strength of the manufacturing sectors in China and the U.S.). The previous decline in Chinese manufacturing activity led the market to follow the stronger AI industry chain-related Hang Seng Technology Index, abandoning the CSI 300, which is more correlated with the overall Chinese economy. However, as the Chinese manufacturing PMI began to recover in February, surpassing the seasonal patterns of 2024 and 2019, this logic may be reversed. On the other hand, the A-H premium has reached a near three-year low. Based on past experience, the recent decline in the Hang Seng Technology Index due to market reflections may open up opportunities for the relative premium recovery of A-shares over H-shares. Currently, the inventory levels of domestic industrial enterprises have fallen below the central level of the past five years, and the operating rate has slightly increased. The implementation of the "two new and two important" policies after the "Two Sessions" will lead to greater marginal improvements in the cyclical sectors of domestic demand. We recommend: first, downstream consumption (brand apparel, food, beverages, white goods, tourism, etc.) that is more related to "volume" stabilization + midstream supply patterns that are relatively good or have positive changes (construction machinery, steel, chemical products, lithium batteries, etc.) + non-ferrous metals (copper, aluminum); second, under inflation risk, commodities are being revalued: gold, crude oil, the initiation of gold stocks may depend on the confirmation of gold price central moving up after the slowdown in gold price increases; third, undervalued + dividends, while also benefiting from the decline in China's macro risks: banks, insurance.

The main text is as follows:

March Transition Arrives as Expected

In last week's report "Expecting the 'March Transition'," we outlined three reasons why a style switch is expected in the market:

First, Chinese technology stocks exhibit characteristics of significant price increases over a defined period, and there has been extreme differentiation in trading heat and volatility among various sectors. Historically, this market characteristic indicates a shorter sustainability of the trend; this week, the total trading volume of A-shares has increased, and the turnover rate continues to rise, with the concentration of trading heat in primary industries still at a high level, and CR5 has further increased, which may imply that although a switch occurred this week, the extent was not sufficient to converge the differentiation.

Second, the trend of technology stocks relies on the continuous emergence of new catalysts. On one hand, even if new catalysts appear, referring to the experience of the fourth quarter of last year, the market's reaction to the catalysts themselves may gradually dull. For example, this Thursday, NVIDIA, a major representative of the AI industry chain, released its Q4 2025 earnings, with revenue and net profit growth rates (78%/80%) exceeding market expectations. However, due to a slight decline in gross margin (75%→73.5%), the market did not respond enthusiastically to this better-than-expected financial report Instead, NVIDIA recorded a 7% decline this week. In this context, the logic that AI-related capital expenditures are driving a broader recovery in investment is less crowded in trading. On the other hand, it should also be considered that new catalysts are difficult to emerge, and the risk of the original narrative being challenged, for example, the "America First Investment Policy" memorandum announced by the Trump administration on February 21 aims to further restrict bilateral investment between the U.S. and China. This memorandum may reflect the U.S. intention to further escalate competition between China and the U.S. across multiple dimensions, including industrial technology.

Third, the momentum of endogenous recovery is gradually becoming evident, with improvements in distribution and social security advancing simultaneously, which will bring opportunities to the long-silent domestic demand sector. These three reasons collectively suggest that investors should at least periodically shift their focus to broader areas beyond AI. Historically, whether it was the mobile internet in 2013 or the high-end manufacturing boom that began in 2019, a single sector is unlikely to dominate the market, and there will be significant rotation in sector performance.

This week (20250224-20250228), the market structure not only saw a "high-low switch" in terms of performance and valuation but also exhibited characteristics of switching from the AI-led technology sector to traditional domestic demand areas. We have compiled the performance and price-to-earnings ratios of popular concept indices from Wind and identified the concept indices with the largest changes in performance rankings compared to last week and since the beginning of the year. The artificial intelligence upstream and downstream industries, such as humanoid robots, computing power, cloud computing, optical modules, and IDC, which have performed outstandingly this year, saw significant declines this week, while sectors related to domestic demand, such as food and beverage, liquor, textiles and apparel, real estate, coal, and retail, experienced a "comeback." The PE (TTM) of these "comeback" sectors is also relatively lower than that of the former.

2 When the "American Exception" Story is Broken

2.1 When the economy weakens, even the capital expenditures of giants cannot be exempted

In this round, whether in the U.S. stock market or the A-share market, investor enthusiasm for the AI sector may stem from expectations of the macro narrative surrounding AI, hoping that AI will become a new engine for economic growth. This engine effect seems to have been validated in the U.S. economic growth over the past two years—between 2023 and 2024, the U.S. real GDP growth rate remains significantly higher than that of other developed economies against the backdrop of maintaining high interest rates.

However, this logic was challenged after the release of several U.S. economic data last week, and it was further shaken after this week's consumer spending and real estate data were published: First, the preliminary value of the U.S. February S&P Services PMI recorded 49.7, hitting a 25-month low, dragging the composite PMI down to 50.4, Creating a new 17-month low, meanwhile, the number of initial jobless claims for the week of February 22 exceeded the levels of the same period in the past three years. As an economy dominated by the service sector, the cooling of the service industry may indicate that the U.S. economy has not escaped its original cycle due to AI. Secondly, the January core PCE data released this week, although further declined and met market expectations, alleviated market anxiety about the Federal Reserve's interest rate hikes triggered by the release of the fourth quarter 2024 core PCE data last week. However, the personal consumption expenditures (PCE) in January showed a month-on-month decline, with the year-on-year growth rate falling for two consecutive months, and the ratio of personal consumption expenditures to disposable income significantly decreased, raising concerns about the weakening momentum of U.S. economic growth. After all, personal consumption expenditures have been one of the largest contributors to U.S. GDP growth over the past two years. Additionally, the National Association of Realtors reported a 4.65% month-on-month decline in the index of existing home sales contracts in January, reaching a historical low, which at least indicates that the traditional part of the economy is still constrained by high interest rates and is currently further affected by rising unemployment and declining purchasing power. Against the backdrop of weakening actual economic data and the various uncertainties brought by Trump’s constant wielding of the "tariff stick," the Atlanta Fed's GDPNow model updated this Friday shows that the U.S. economy will shrink by 1.5% in Q1 2025, marking the first negative growth since Q1 2022.

In this context, even investors who are fully confident in AI's ability to enhance productivity should currently consider a risk: If developed economies represented by the United States enter a recession, will it affect the capital expenditure and prosperity of the AI industry? **In our previous report "When AI Becomes the Focus," we compiled the capital expenditure figures of the seven major U.S. tech giants. Their high growth rate and substantial investments have indeed driven U.S. capital expenditure since Q2 2023. However, can this be sustained? Although technological breakthroughs and prospects are the initial driving forces behind capital expenditure, investment activities are always constrained by two main factors: on one hand, the ability of their traditional core businesses to generate cash flow, which is closely related to overall demand; on the other hand, whether the financing environment is favorable, which is related to the Federal Reserve's monetary policy.Looking at the present, the capital expenditure of the seven giants accounted for a historical high of 40.5% of their operating cash flow net value in Q1 2023. Subsequently, due to a slowdown in the growth rate of operating cash flow net value and a decline in the interest rate hike cycle, this figure dropped, but is expected to rebound significantly to a high level of 39.5% in 2024. Assuming this ratio remains unchanged, and the growth rate of the seven giants' operating cash flow net value has shown a significant negative correlation with the inversion of U.S. Treasury yields (2Y-10Y) over the past four years, the current difference between the 2-year and 10-year U.S. Treasury yields has rebounded, and recession expectations are growing. Therefore, capital expenditure may also face challenges due to the slowdown in operating cash flow growth, especially in the context where emerging AI-related businesses have yet to find a clear path to profitability.

2.2 The risk of inflation has not retreated, and overseas liquidity remains tight

In our report before the Spring Festival, we pointed out that the implementation of Trump's tariff policy in the coming quarter is an important indicator to observe his policy priorities, which will largely affect inflation levels and constrain the Federal Reserve's monetary policy space. After announcing a 10% tariff on Chinese products and a 25% tariff on Canadian and Mexican products on February 1, the Trump administration continued to wield the "tariff stick." On February 27, it announced that tariffs on Canada and Mexico would be implemented starting March 4, along with an additional 10% tariff on China. Under the premise of a weakening domestic economy in the U.S. and a controllable inflation level, the Trump administration further elevated the priority of the "tariff card." However, its recent efforts to reduce inflation on the supply side have made little progress: For example, the slow progress of peace talks between Russia and Ukraine, with the March 1 meeting with Zelensky yielding no good results; the "productive investment" support for the U.S. domestic market mentioned by Treasury Secretary Mnuchin has yet to show any results; recently, the Trump administration's focus has not indicated that it is working hard to pressure OPEC+ to increase oil production to achieve the goal of reducing inflation. **The current U.S. inflation seems to be more aligned with a decline in demand rather than a reduction in supply-side uncertainty due to rising tariffs, ultimately pointing to either a weakening of demand Either we will see a rebound in inflation, ** which also explains the "hawkish" remarks made by Richmond Fed President Barkin in the middle of this week, stating that "in the hope of ending the battle against inflation, all uncertainties require the Federal Reserve to remain cautious with interest rates."

2.3 A divergence begins to appear for Chinese tech stocks

As signs of a weakening U.S. economy and marginal tightening of overseas liquidity emerge, and flaws at the industrial level appear in overseas companies, can Chinese tech stocks enjoy a premium and become a safe haven for funds investing in technology? We believe that a divergence will also begin to appear within Chinese tech stocks: On one hand, from the perspective of overseas mapping, after Nvidia's earnings release, its valuation has been digested to 41.8 times PE. As the core of the current global AI industry chain, its valuation has a certain "anchoring effect," especially after Deepseek emerged to challenge its monopoly position, this valuation level can be seen as a ceiling for some time. The general rise of Chinese tech stocks is constrained by overseas mapping targets; on the other hand, within tech stocks, the realization of performance will become a watershed for the market. Historically, based on typical industry cycle experiences, A-share investors' early moves in the industry cycle often do not exceed one year. In contrast, the realization of performance elasticity in the early-moving combinations over the past two years has been somewhat "long overdue."

3 Chinese cyclical assets are not just low-position defenses

3.1 The diffusion market has just begun

Since the beginning of this year, the performance of Hang Seng Technology has significantly outperformed the cyclical A-shares represented by the CSI 300, possibly due to: the U.S. is shifting from manufacturing to services, with the key driver being the AI technology wave it leads. China's economic foundation is manufacturing, and after the second half of 2021, the importance of manufacturing has been further enhanced through "de-real estate and financialization." However, since the beginning of this year, the U.S. manufacturing PMI has risen more than China's manufacturing PMI, which has led the market to choose to follow the stronger AI industry chain-related Hang Seng Technology, abandoning the CSI 300 that is more related to the Chinese economy. As China's manufacturing PMI began to recover in February, surpassing the seasonal patterns of 2024 and 2019, this may reverse the aforementioned logic.

On the other hand, the A-H premium has reached a near three-year low. Based on past experiences, the recent decline of Hang Seng Technology due to mapping may open up an opportunity for the relative A-share to H-share premium recovery.

3.2 The Tailwind of Domestic Demand is Accumulating

After the "9.24" event last year, the domestic demand economy experienced a continuous recovery for four months (except for January, when high-frequency data on the production side declined due to the Spring Festival). The sustainability of this recovery momentum will become the focus of market attention for some time to come. Next week, the highly anticipated "Two Sessions" will be held, which will also lead the market's focus towards the fundamentals. Currently, the inventory levels of domestic industrial enterprises have decreased, only higher than those in 2023 and 2019 during the same period, and the construction-related operating rates have increased. The implementation of the "dual 重双新" policy will lead to greater marginal improvements in the domestic demand-related cyclical sectors compared to external demand and overseas reflections. We believe that the current layout in cyclical sectors can start from downstream consumption, move to the midstream manufacturing sector where supply is clearing, and then to the upstream physical asset sector driven by demand.

From the February PMI, "volume increase" remains the main tone of the current economic recovery, with production recovery outpacing order recovery, and the rise in raw material prices exceeding the rise in factory prices. On this basis, the cyclical sectors that should benefit are those more related to "volume" rather than "price." We reviewed the correlation of 108 secondary industries in the A-share market with electricity consumption and PPI, finding 43 industries more related to electricity consumption, mainly in downstream consumption (beverages, white goods, home furnishings, catering and tourism), midstream manufacturing (electrical equipment, general equipment, textile manufacturing), and upstream materials (industrial metals, paper making, etc.).

4 Shift to Cyclical Stocks

After the self-adjusting market following the extreme AI market trend, a macro scenario has emerged that is more favorable for Chinese assets. The continuous improvement of PMI and the upcoming "two new and two heavy" policies are accumulating the power to repair the AH premium and the relative difference between the CSI 300 and the Hang Seng Technology Index.

We recommend: First, in the context of gradually improving domestic fundamentals leading to better expectations, domestic cyclical-related midstream manufacturing (engineering machinery, steel, chemical products, lithium batteries, etc.) + consumption (branded clothing, food, beverages, white goods, tourism, etc.) + non-ferrous metals(copper, aluminum)**;

Second, under the logic of physical assets, nominal interest rates will lag behind inflation, and globally priced, dollar-denominated commodities will continue to be revalued: gold, crude oil, the initiation of gold stocks may depend on the confirmation of the upward shift in gold prices after the pace of gold price increases slows down;

Third, undervaluation + dividends, while also benefiting from the decline in China's macro risks: banks, insurance.

Authors: Mu Yiling, Wang Kuangwei, Source: Yiling Strategy Research, Original Title: "March Transition Begins 丨 Minsheng Strategy"

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