CICC: Have US stocks fallen into place?
The Zhitong Finance App learned that CICC released a research report saying that recent fluctuations in US stocks have clearly increased, and policy “noise” after the midterm elections has induced market turmoil. The key point is that AI itself is currently facing many “problems” such as “no response” in demand and high investment, which was revealed early. Of course, this rapid clearance may not be bad for long-term trends; it helps absorb excessive expectations and positions. This is beneficial for regaining upward momentum in the future with the emergence of new catalysts; it is only necessary to wait for new catalysts to emerge. You need to know that the increase in US stocks since this year has basically been contributed by profit rather than by valuation expansion. After a recent correction, valuations have declined even more rapidly. With the exception of the Dow Jones 19.6 times valuation, which is still 86% of the historical quantile over the past 20 years, the S&P 500 and NASDAQ valuations fell back to 17.8 times and 20.3 times, respectively, corresponding to 70% and 44% of the past 20 quantiles. Cost half even dropped to 13.8 times, which is only 14% of the past 20 years quantile.
CICC's main views are as follows:
The “K-shaped differentiation” of US fundamentals this year has concentrated the market and relied on a few directions: 1) US Mag7 contributed 22% of the S&P 500 increase since 2026, and contributed 40% to the previous decline in mid-October; 2) the profits of the top five cloud vendors accounted for 20% of the S&P 500 index in the first three quarters of 2026, contributing 25% of the S&P 500 profit growth; 3) the leading concentration effect is still the same Yes, the S&P 500 has risen 8% since the beginning of the year, but the share ratio on the 200-day EMA fell to 55% from 60% at the beginning of the year. The divergence was particularly obvious in August-September. However, AI currently lacks models and application-level catalysts, and is also facing policy uncertainty, and traditional demand is being suppressed by high interest rates. In other words, the main direction is facing challenges, and the traditional direction has yet to be taken over, which explains the overall turmoil.
Therefore, the market needs two catalysts: either a huge breakthrough in AI, or an end to austerity or even easing by the Federal Reserve. Until then, CICC suggests watching carefully: 1) focus on whether there is negative feedback on liquidity in the short term, or further policy disturbances (game before the new National Assembly takes office early next year); 2) control positions to reserve space for subsequent allocations; 3) adjust the structure. Even within technology, priority can be given to cloud vendors and computing chips with low AI exposure or low valuations and more bargaining power.
The first quarter of next year is a key point: 1) The opportunity for the Federal Reserve to clearly shift is expected until the beginning of next year. Strong non-agricultural agriculture in October strengthened the probability of interest rate hikes in December, and it is more certain that inflation will fall back early next year, driven by a high base; 2) The quarterly report to be disclosed in January next year. Technology companies' ROIC may continue to be under pressure, and AI will be even more needed to implement new developments at that time; 3) The new US Congress will take office on January 3 next year. Trump's pre-term policies and newly elected lawmakers may continue to be disrupted. The market needs to be digested, and the new Congress may gradually desensitize after taking office. So, at that point, either AI will have to make progress, or the Federal Reserve's austerity expectations will decline, which may provide support or opportunity.
In this article, CICC further combines the third-quarter results of US stocks, which have already been basically disclosed, to answer questions about the outlook for US stocks from a fundamental perspective.
Market driving force: profit dominates the market, AI contributed 60% to growth; the growth rate reached an inflection point
First, strong earnings growth is the main driving force for US stocks. Since the beginning of the year, major index gains in US stocks have all been contributed by profit. The S&P 500, Dow Jones, and Nasdaq indices have risen 8%, 7%, and 6%, respectively. Among them, dynamic profit has increased by 29%, 17%, and 40%, respectively, offsetting the contraction in valuation. S&P 500 profit increased 25% in the third quarter, down from 32% and 56% in the first quarter and the second quarter, but still higher than the average quarterly profit growth rate of 12% for the 2023-2025 period.
Second, the AI-related sector contributed 60% to profit growth. Using communications services, information technology, and Amazon to estimate the profit situation in the AI sector, the third quarter contributed 62% of overall profit growth, down from 70% in the second quarter. Among them, information technology contributed 58%, supporting the vast majority of growth in the AI sector, while communications services clearly weakened due to a sharp slowdown in the growth rate of media and entertainment in the industry. This also makes overall profit growth tied to AI. If the AI sector is excluded, the profit growth rate may fall back to 15% in the third quarter.
The profit of traditional demand in the third quarter was not bad, but the suppression of high interest rates still dragged down profits. Looking at the growth rate, sectors such as energy (150%), semiconductor equipment (110%), real estate management and development (49%), and capital goods (48%) are leading, while transportation (-37%), telecommunications services (-34%), financial services (-14%), and insurance (-11%) lag behind. In terms of contributions, sectors such as industry, raw materials, utilities, and essential consumption contributed 10% of profits, while financial and real estate contributions all turned negative, dragging down 7.5% of overall growth.
Third, the growth rate reached an inflection point, decelerating for the first time since 2Q25. S&P 500 earnings increased 25% in the third quarter, down from the first two quarters of this year. The market is more “picky” about performance. Not only does it require profits to maintain a high growth rate, but it also needs to continuously exceed expectations and accelerate, making it more difficult for the third quarter earnings to exceed expectations. The ratio and magnitude of the S&P 500 index constituent stocks fell slightly from 86% in the second quarter to 84%, while the margin that exceeded expectations fell significantly from 27% in the second quarter to 3.5%, the lowest level since the fourth quarter of 2022. Among them, Google and Meta's communication services declined the most, with profits exceeding expectations by 99% in the second quarter to falling 19% in the third quarter.
AI direction: income lags behind investment; returns are lower, but still higher than costs; can withstand interest rates falling to 6-6.5%
1) Revenue lags behind investment. Capital expenditure is expanding at an accelerated pace, free cash flow is negative, and pressure on external financing is rising. The total capital expenditure of the five major cloud vendors rose from US$181.5 billion to US$230 billion in the third quarter, and the year-on-year growth rate accelerated from 87% in the second quarter to 118%. However, the growth rate of operating cash flow declined markedly, from 38% in the second quarter to 19% in the third quarter. The five major cloud vendors totaled 19.8 billion US dollars, which was insufficient to cover capital expenses during the same period. As a result, the free cash flow of the five major cloud vendors changed from US$4.8 billion in the second quarter to US$31.8 billion. With the exception of Microsoft, the free cash flow of Amazon, Meta, Google, and Oracle all turned negative in the third quarter, further boosting external financing needs. Against the backdrop that demand for AI has not broken through the “ceiling” of existing application scenarios and the prospects for return on investment are still uncertain, market attention on financing costs and funding sources has increased markedly, and concerns about credit risk have intensified, and CDS of the five major cloud vendors has risen to historic highs.
2) The return is higher than the cost. Marginal ROIC declined further, but overall returns can still cover financing costs. As investment capital continued to expand, new operating returns were not synchronized, and the marginal ROIC of the five major cloud vendors continued to drop to 6.1% from a high of 44% in the second quarter of 2024, driving the overall ROIC further back to 15%. However, the overall ROIC is still higher than the current weighted average cost of capital (WACC) of about 10% of the five major cloud vendors, which means that cloud vendors can still get excessive returns.
The decline in marginal returns has reduced the interest rate level that AI can withstand to 6-6.5%. Previously, CICC estimated that if the five major cloud vendors kept their ROIC unchanged, assuming that the difference between ROIC and WACC narrowed to the average level of 8ppt before the rise of the AI trend in 2015-2023, the corresponding bearable risk-free interest rate would be 7 to 7.5% (“AI from “friend” to “enemy” of traditional requirements?). However, as the marginal return on investment weakened further in the third quarter, the difference between ROIC and WACC has converged to a historical average. Combined with return expectations for 2027, the risk-free interest rate level that AI can withstand fell to 6-6.5%.
3) The “capitalization” of cloud vendors puts pressure on capital turnover. ROIC was further split into after-tax operating profit margin (noPat/revenue) and invested capital turnover (revenue/IC). The operating profit margin of the five major cloud companies after tax has continued to rise since 2023, from 14.8% to 22.1% in the third quarter; however, the capital turnover ratio continued to decline, falling from a high of 109% in the first quarter of 2023 to 66% in the third quarter, a new low since 2015. At the same time, since the second quarter of 2025, the slowdown in profit margin improvement has not been able to offset the drag on capital turnover, which has also led to the continued decline in ROIC of the five major cloud vendors.
There is further differentiation within technology, and some hardware aspects still need to absorb growth expectations, and cloud manufacturers agree on prices that are relatively reasonable. Some hardware infrastructure that takes into account high growth expectations, such as network interconnection/optical modules, semiconductor manufacturing, liquid cooling/temperature control, etc. Recently, valuations in the semiconductor manufacturing sector have declined rapidly to around 20% from the 90% quantile in early October. In contrast, cloud vendors did not give significant valuation premiums to new profits, and there is relatively limited room for valuation adjustments.
Traditional sector: Interest rate sensitive links are being suppressed by high interest rates, and the AI squeeze is gradually reflected
Traditional demand has yet to be fully improved. In terms of consumption, consumption of services is superior to consumption of goods. Consumer services, food and beverage and tobacco, and automotive and parts profits increased by 12%, 8.7%, and 7.3%, respectively, while durable goods and clothing, and household and personal goods declined by 5.4% and 10.3%, respectively. The industrial sector is more divided. Profits in capital goods increased by 48.3%, commercial and professional services increased by 16.2%, but transportation, which was affected by high oil prices, declined by 37.1%. In the financial sector, bank profits increased slightly by 4.8%, supported by net interest income, while insurance and financial services declined by 10.7% and 13.6%, respectively, due to factors such as fluctuations in investment profit and loss.
Growth is still resilient, but interest rate sensitive growth links continue to bottom out. Recently released macro data shows that US growth is still resilient. Real GDP in the third quarter grew by 3.4% month-on-month annualized, higher than expectations of 3.2% and 2.2% in the second quarter. In addition to AI-related investments, private consumption is also strong, especially consumption of services such as medical care, catering, and lodging. ISM manufacturing and non-agricultural industries also surpassed expectations in October, and the unemployment rate also fell from 4.2% to 4.1%. However, high interest rates continue to curb residents' demand for housing purchases and credit. Interest rates on 30-year mortgages have continued to rise since the beginning of the year, and once rose to 7.4% in October. Real estate sales remained low. Although existing home sales rebounded slightly month-on-month in September, annualized sales were only 3.99 million units, and new home sales also fell to 676,000 units. On the other hand, AI with weakening returns may also begin to “crowd out” the physical and financial factor resources of traditional demand. As CICC emphasized in “From the “Friend” to “Enemy” of Traditional Demand, the positive externalities brought about by AI demand previously supported US economic growth, but as AI required more external capital to support investment, the crowding out effect on traditional demand was also unavoidable. Massive AI debt issuance and the Federal Reserve's need to raise interest rates are typical examples.
Larger repairs will still need to be matched by falling interest rates. CICC previously estimated that the interest rate on 10-year US Treasury bonds that traditional demand can withstand is about 4% using real estate and non-AI companies. Despite current resilience, high interest rates continue to squeeze residents' housing purchases and the further release of credit demand from small and medium-sized enterprises, which is reflected in insufficient momentum for credit expansion in the residential sector. In the short term, growth resilience and employment data make the Fed lack the momentum to make an immediate shift, but weak traditional demand does not support the Fed's larger rate hike. If AI expectations continue to weaken in the first quarter of next year, then the market's austerity expectations for the Federal Reserve are expected to decline, creating conditions for falling interest rates and restoration of traditional demand.
Prospective judgment: The market needs to end austerity or catalyze AI progress. The beginning of next year is a key point
Currently, the market's expectations for AI itself are in a “no match” stage. The key problem is that there are no new narratives and new catalysts for demand and models, so it will also amplify the risk of credit and regulatory pressure. Intense capital expenditure continues to squeeze cloud vendor's free cash flow, leading to growing concerns about the speed of AI revenue payout and return on investment. This is similar to the logic that the market worried about private equity and credit risks when AI progress was insufficient at the end of last year.
In this context, the market needs two catalysts: either a huge breakthrough in AI, or an end to austerity or even easing by the Federal Reserve. If none of them appear, it may affect investors' willingness to return to the bottom, and it will also amplify fluctuations. In the short term, we need to pay attention to whether there is negative feedback on liquidity or further policy disturbances (whether Trump will use this window to play policy games before the new Congress takes office early next year). The first quarter of next year is a key verification point. There will be changes when the National Assembly arrives, AI is realized during the performance period, and inflation falls back.
Until then, CICC still advises to wait and see carefully in the short term. Until the above catalyst appears, focus on whether negative feedback on liquidity and further policy disturbances will occur in the short term. At present, the S&P 500 index has returned to close to the annual line, and early adjustments have released some sentiment, but the market still lacks new upward catalyst. Factors such as liquidity segments and the December Federal Reserve's interest rate hike may also cause further disturbances. Therefore, CICC suggests waiting for the above catalysts to appear and the return on investment prospects are more clear before allocating. Looking at the extreme situation, if all of the additional profit growth expected by AI is “erased”, and the dynamic profit growth rate may fall back to 10%. Assuming that valuations fall back to 16 to 17 times at the same time, there is still room for about 10% decline in the corresponding S&P 500 index.
In addition to controlling positions, it is also possible to adjust the position structure. Even within technology, priority can be given to cloud vendors and computing chips with low AI exposure, or low valuations, and more bargaining power. Recent market performance also reflects this difference. The pullback was more obvious when previous growth expectations were fully factored in, while cloud vendors were relatively resistant to decline. Using the average valuation over the past 10 years as a reference, sectors such as optical modules, semiconductor manufacturing, power grids, and distribution still need to raise profits by 40 to 70% to match current pricing, while cloud vendor profits have room for a profit reduction of nearly 20%.