Sinolink: Tech stocks saw significant volatility this week, and current conditions are not yet suitable for bottom-fishing.

Sinolink released a research report stating that this week, the investment returns of overseas tech giants were once again scrutinized by the market, and tech stocks saw significant volatility. Comparing with the 2007-2008 cyclical stock cycle and market trading indicators, based on historical extreme market conditions, current conditions are not yet suitable for bottom-fishing. If trading indicators further reach extreme values and coincide with stabilization in US stocks, that may be the timing for a phased rebound. Sinolink recommends: First, the energy chain represented by coal, oil, and oil shipping will continue to benefit from global restocking demand for energy; the energy chain is a resilient asset under capital cost volatility and geopolitical conflicts. Second, from an earnings perspective, recommend focusing on: Continued high prosperity: medical services, specialty retail chains, coal mining, shipping ports, as well as sub-sectors within non-ferrous metals and chemicals; Turnaround 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 the turning point of real interest rates, but referencing the 2008 experience, major opportunities are also being incubated as a result. The emergence of an inflection point: Investment returns of overseas tech giants are once again being scrutinized by the market From this week's US cloud vendors' earnings reports, AI-related investment remains persistently high, but the speed of realizing profits and cash flow still struggles to match the expansion of capital expenditures, and the sustainability of future investment returns and financing has not received clear guarantees. The continued growth of current cloud business revenue means that AI investment is being converted into quantifiable investment returns, in line with expectations; but at the same time, the contradiction between Shanghai New World and the Old World is also intensifyingGoogle's advertising and Amazon's e-commerce revenue growth rates are both marginally declining. While the development of new businesses is not yet sufficient to bring abundant cash flow, the weakening growth momentum of traditional businesses has undermined cloud vendors' ability to support massive capital expenditures through internal cash flow. From the perspective of debt and revenue, as cloud vendors' external financing scale expands, interest expenses have also begun to erode profits to some extent. Typically, Amazon's interest expense as a share of net profit rose from a 2025 average of 3% to 8.2% in 26Q3. Although new businesses are indeed growing, damage to traditional businesses combined with a heavier interest burden has instead caused Google's and Amazon's free cash flow to fall sharply, which also means that cloud vendors currently still struggle to form a positive cycle of "cloud business growth cash flow recovery capital expenditure expansion." After the earnings disclosures, different performance among cloud vendors 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 raised and its stock price rising after the earnings disclosure; while Amazon and Google saw sharp stock price declines due to earnings misses (EPS lowered after earnings disclosure) and plunging cash flow. Currently, for US Treasury yields, with the consensus forecast of AI capital expenditure slowing in 2026Q4, the peak in US Treasury yields may have already appeared, but considering that the current US economic fundamentals remain resilient, the future downward magnitude of US Treasury yields still needs to be observed, and a rapid decline in rates still requires waiting for the next signpost signal of a rapid contraction in AI financing. Observing the magnitude of the tech correction: 2007 cyclical stocks vs. 2026 hardware tech Observing the magnitude of the tech correction from two dimensions: (1) Sinolink's previous mid-year strategy "Until the End" proposed that this round of AI industry cycle can be analogized to China's manufacturing investment cycle in 2005-2007. Comparing the stock price drawdowns of cyclical stocks in 2007-2008 with this round of tech stocks, there are similarities: this round of tech stocks has seen a drawdown of about 54% since the correction began in early July, occurring before US tech stocks had only reversed their uptrend but had not yet seen an obvious correction; in 2007-2008, before the US stock market peaked and before the subprime crisis, Chinese cyclical 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 US stock market decline in 2008 occurred after August 2008, corresponding to the escalation of subprime crisis risks and simultaneous declines in both volume and price across the industrial chain; at the same time, A-shares were not spared eitherafter August 2008, the All-A index still had a decline of more than 40%, and related cyclical sectors fell by nearly 50%. At present, China's AI hardware-related assets are upstream suppliers to the US, similar to how China was a supplier to US demand after joining the WTO in 2007, and subsequent developments in the US tech and debt cycles need to be closely monitored. Currently, although there are still concerns about the sustainability of the cloud vendor model, the AI commercialization path is indeed materializing, and we remain in a period of divergence. (2) From market trading indicators: The margin financing guarantee ratio continued to decline this week; Implied volatility of major broad-based indices is risingcurrently the implied volatility of the STAR 50, ChiNext Index, and CSI 1000 has all risen to above the 90th percentile level over the past 3 years, but the above two indicators still have some distance from historical extremes. From the perspective 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 moving toward historical extremes could still bring downside 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 are not yet suitable for bottom-fishing. If trading indicators further reach extreme values and coincide with stabilization in US stocks, that may be the timing for a phased rebound. During periods of declining earnings: Structural clues for coping with market volatility The current cyclical slowdown in corporate earnings has begun to be reflected in economic data and corporate earnings reports: since June this year, the growth rate of industrial enterprise profits has marginally declined for four consecutive months, and at the same time, listed companies' profit side began to show signs of cooling in 2026Q3. Structurally, earnings 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 earnings in non-tech sectors have slowed, the prosperity resilience of some industries remains. Rather than repeatedly speculating amid the high volatility of the tech sector, it is better to expand one's horizon to non-tech sectors. From an earnings perspective, one can focus on the continued high-prosperity combination (12 industries in total, such as coal mining) and the turnaround reversal combination (7 industries in total, such as refining and trading). Accumulating strength for future market conditions Although value investors should "be greedy when others are fearful," judging when true "fear" occurs is the greatest difficulty. As Chinese investors participating in this round of global industrial chain mapping, the decline magnitude of the domestic market itself is not the best measure of "fear"this is also the biggest trap in the bottom-fishing process. After the inflection point in US tech giants' capital expenditures appears, Sinolink is full of confidence in the decline of capital costs over the next 1-2 years, the blossoming of tech applications, and the rise of Chinese assets, and is now committed to waiting for that day to arrive.