Sinolink Securities Co., Ltd. (600109) has released a research report stating that this week the investment returns of overseas tech giants have once again come under market scrutiny, with tech stocks experiencing significant volatility. Comparing the 2007-2008 cycle stocks and market trading indicators, based on historical extreme market conditions, current conditions do not meet the criteria for bottom-fishing. Only if trading indicators further reach extreme values and coincide with a stabilization in U.S. stocks might that signal a timing for a phased rebound.
Sinolink Securities recommends: First, the energy chain represented by coal, oil, and oil shipping will continue to benefit from global energy restocking demand; the energy chain is a resilient asset amid capital cost fluctuations and geopolitical conflicts. Second, from an earnings perspective, attention should be paid to: ① Continued high prosperity: medical services, professional chain stores, coal mining, shipping ports, as well as sub-sectors within non-ferrous metals and chemicals; ② Distressed reversal strategy: sub-sectors within textiles and apparel, as well as papermaking, refining and trading, and medical devices. Third, gold may face liquidity shocks during periods when real interest rates turn, but referencing the 2008 experience, major opportunities are also being incubated as a result.
Emergence of the Turning Point: Investment Returns of Overseas Tech Giants Come Under Scrutiny Again
From this week's earnings reports of U.S. cloud providers, AI-related investments remain persistently high, but the pace at which profits and cash flows are being realized still struggles to match the expansion of capital expenditures, and the sustainability of future investment returns and financing has not obtained clear guarantees. The continued growth of cloud business revenue means AI investments are being converted into quantifiable investment returns, which is in line with expectations; however, at the same time, the contradiction between the new world and the old world is intensifying, with Google's advertising and Amazon's e-commerce revenue growth rates both declining marginally. While the development of new businesses is not yet sufficient to bring in ample cash flow, the weakening growth momentum of traditional businesses has undermined the ability of cloud providers to rely on internal cash flow to support massive capital expenditures. From the perspective of debt and revenue, as cloud providers expand external financing, interest expenses are beginning to erode profits to some extent. A typical example is Amazon, where interest expenses as a share of net profit rose from an average of 3% in 2025 to 8.2% in 2026Q3. Although new businesses are indeed growing, the damage to traditional businesses combined with heavier interest burdens has caused Google's and Amazon's free cash flows to decline sharply, meaning cloud providers still struggle to form a positive cycle of "cloud business growth → cash flow recovery → capital expenditure expansion." After earnings disclosures, the different performance of each cloud provider has also led to differences in stock price performance. Microsoft was "rewarded" by the market for being able to achieve positive free cash flow, with EPS revised upward and share price rising after the earnings disclosure; while Amazon and Google saw sharp stock price declines due to earnings misses (EPS revised downward after earnings disclosure) and plummeting cash flows.
Currently, regarding U.S. Treasury yields, with the consensus expectation of AI capital expenditure slowdown in 2026Q4, the peak in U.S. Treasury yields may have already appeared. However, considering that the current U.S. economic fundamentals still possess resilience, the extent of future declines in U.S. Treasury yields remains to be observed, and a rapid decline in rates still requires waiting for the next landmark signal of rapid contraction in AI financing.
Observing the Magnitude of Tech Adjustment: 2007 Cycle Stocks vs. 2026 Hardware Tech
The magnitude of the tech adjustment can be examined from two dimensions: (1) Sinolink Securities' earlier mid-term strategy "Until the End" proposed that this AI industry cycle can be analogized to China's manufacturing investment cycle from 2005-2007. Comparing the stock price drawdowns of cycle stocks in 2007-2008 with this round of tech stocks reveals similarities: this round of tech stocks has seen a drawdown of approximately 54% since the adjustment began in early July, occurring before U.S. tech stocks had only seen an upward reversal without a significant adjustment; before the 2007-2008 U.S. stock market peak and the subprime crisis, Chinese cycle stocks peaked in October 2007 and saw a drawdown of 62% by June 2008. However, it is worth noting that the real acceleration of the U.S. stock market decline in 2008 occurred after August 2008, corresponding to the escalation of subprime crisis risks and simultaneous declines in volume and price across the industrial chain; at the same time, A-shares were not spared either, with the all-A index still falling more than 40% after August 2008, and related cyclical sectors also declining by nearly 50%. Currently, China's AI hardware-related assets are upstream suppliers to the U.S., similar to how China became a supplier to U.S. demand after joining the WTO in 2007, and subsequent developments require close attention to the U.S. tech and debt cycle conditions. At present, although there are still concerns about the sustainability of the cloud provider model, the AI commercialization path is indeed manifesting, and we remain in a period of divergence. (2) From market trading indicators: ① The margin financing guarantee ratio continued to decline this week; ② Implied volatility of major broad-based indices is rising, with the implied volatility of the STAR 50, ChiNext Index, and CSI 1000 all having risen to above the 90th percentile over three years, but the above two indicators still have some distance from historical extremes. ③ In terms of market breadth, both the overall A-share market and the internals of tech stocks have fallen to historical lows. The current trading structure has indeed begun to meet bottom conditions, but the fact that trading indicators still have some distance from historical extremes means that if tail risks emerge, related indicators converging toward historical extremes could still bring downward risks to the market. It is worth mentioning that the current overweight ratio in AI hardware remains higher than the peak overweight ratios of most historical mainstream sectors, and actively managed equity funds' overall position exposure to AI hardware remains relatively high.
The conclusion is that, based on historical extreme market conditions, current conditions do not meet the criteria for bottom-fishing. Only if trading indicators further reach extreme values and coincide with a stabilization in U.S. stocks might that signal a timing for a phased rebound.
During Periods of Declining Profits: Structural Clues for Navigating Market Volatility
The current cyclical slowdown in corporate profits is beginning to be reflected in economic data and corporate earnings reports: since June of this year, industrial enterprise profit growth has declined marginally for four consecutive months, while listed companies' profit levels began showing signs of cooling in 2026Q3. Structurally, the profit diffusion measured by the proportion of positive profit growth has been higher only than 22Q4 since 2021, prosperity remains scarce, and profit growth is mainly concentrated in a few high-prosperity industries. It is worth noting that although overall profits in non-tech sectors have slowed, the prosperity resilience of some industries remains. Rather than repeatedly gambling amid the high volatility of the tech sector, it is better to expand one's horizons to non-tech sectors. From an earnings perspective, attention can be paid to the high-prosperity continuation portfolio (12 industries such as coal mining) and the distressed reversal portfolio (7 industries such as refining and trading).
Building Strength for Future Market Conditions
Although value investors should "be greedy when others are fearful," judging when true "fear" has arrived is the greatest difficulty. As Chinese investors participating in this round of global industrial chain mapping, the magnitude of decline in the domestic market itself is not the best measure of "fear," and this is also the biggest trap in the bottom-fishing process. After the inflection point in capital expenditures of U.S. tech giants appears, Sinolink Securities is full of confidence in the decline of capital costs, the blossoming of tech applications, and the rise of Chinese assets over the next 1-2 years, and will now strive to wait for that day to arrive.
Comments