Global technology stock volatility: How to understand? How to respond?

1. How to understand the recent fluctuations in global technology stocks? This week, there has been another increase in volatility in global technology stocks. One of the main reasons behind this is the crowded chip structure, especially the extreme leverage in the South Korean stock market which has amplified the volatility. However, in fact, the AI sector has not experienced a general decline recently, but rather a "shrinking circle" focusing on the storage industry chain that was validated first by Micron's performance announcement (storage chips, semiconductor material equipment). Therefore, we believe that the recent fluctuations in the technology sector are more important due to the fact that both China and the US are currently in a period of no significant performance announcements, leading to a lack of clear fundamental guidance in the market. Funds can only speculate around changes in the industry margins, some of which may be considered as "noise." Some "ghost stories" that have led to the recent increase in volatility in global technology stocks: 1) The accelerated pace of bond issuance by North American cloud companies and changes in their financing structure have raised concerns in the market that their own cash flow may not be enough to cover their continuous expansion of capital expenditure. Coupled with the recent decline in token usage, large technology companies led by MAG7 have become a drag on the Nasdaq's decline. 2) Rising storage costs have forced Apple and Microsoft to raise prices of their products, further exacerbating concerns about the sustainability of high capital expenditure after the demand side is killed by the cost of computational power. Therefore, both global "computing power users" and "computing power producers" have experienced declines. 3) OpenAI announced a delay in going public, causing concerns in the market that top star companies' operational data may not meet expectations, leading to a cautious investment trend across the entire industry. The core concerns of the market mentioned above all revolve around the key indicator of "Hyperscalers' capital expenditure growth rate" (Hyperscalers - Microsoft + Google + Amazon + Meta + Oracle, see full text). Looking back, it is indeed a more straightforward, effective, and forward-looking indicator for tracking and predicting trends in the AI industry. Regardless of hardware/software stocks in the US stock market, the expected year-on-year change over the next 12 months is strongly correlated with Hyperscalers' capital expenditure year-on-year. However, there are many factors that can affect capital expenditure, and there is a lot of data and changes that can be considered as "noise." Ultimately, it is necessary to focus on the main contradictions: fundamentally, the change in Capex is mainly determined by the return on investment, and the return on investment is currently linked to the ARR of top large-scale model manufacturers. This data will be more transparent after Anthropic's IPO, supporting the continued rise in ROI and Capex. As long as computational power can still improve AI performance and the ROI generated by tokens can still be maintained, top model manufacturers will continue to increase their capital investment. Cloud service providers, as the "computational power suppliers," receive indirect benefits that are enough to support further increases in capital expenditure. The market expects North American cloud companies their free cash flow will turn negative this year, questioning the sustainability of capital expenditure. However, what is more important is the growth of future operating cash flow, which can support high capital expenditure growth. As long as the growth in operating cash flow brought by cloud and AI businesses can exceed the growth in capital expenditure, future free cash flow will also turn positive, corresponding to sustainable high capital expenditure. Since the beginning of this year, long-term revenue expectations of major North American cloud companies have continued to rise, and the upcoming earnings season in mid to late July will be a key window to verify whether the revenue from AI business can support their high capital expenditure growth. Finally, based on historical experience, the earnings season will be an important window for the upward revision of capital expenditure expectations. From 2024 to 2026, the initial capital expenditure growth rate was approximately three times lower than the final figure. Whenever listed companies disclose new guidance during the earnings season, it is an important window for the market's expectation of an upward revision in Capex. Therefore, the recent volatility in global technology stocks is partly attributable to the amplification effect of crowded structures and extreme leverage, but fundamentally it is due to the lack of clear fundamental guidance during the period of no significant performance announcements, leading to a game of long and short positions based on global capital expenditure expectations and changes in industry margins. However, the fundamental situation in the technology industry has not undergone many substantial changes. The upcoming earnings season will determine the true picture of the technology industry, and the recent discrepancies and volatility caused by some "noise" may actually create opportunities for positioning. 2. In China, the root cause of recent structural differentiation - economic conditions, will face another important validation window For A-shares, in addition to the changes in the global AI industry, we have emphasized before that the recent structural differentiation in the market is fundamentally caused by the differentiation of domestic economic conditions, and the upcoming early July period of earnings forecasts will be another validation of this differentiation. This means that the consensus on growth dominance and economic conditions will not change due to recent fluctuations. Currently, whether from a macro, meso, or micro perspective, all indicators point to a further widening of the economic differentiation between emerging industries and traditional industries in the mid-year report: 1) Macro: In May, both social retail sales and investment growth rates turned negative, while high-tech industries related to AI and export growth rates continue to rise. 2) Meso: TMT related to AI, upstream resource products (mostly non-ferrous metals, chemicals), industrial enterprise profit growth has been on the rise since the beginning of this year, while consumer, midstream manufacturing, public utilities, and other industries have seen a continued downturn. 3) Micro: After the first quarter, only TMT hardware and upstream resource products have seen upward revisions in 2026 net profit estimates, while other industries have experienced downward revisions. Therefore, as the market's focus in July shifts to industry trends and earnings verification, the consensus on growth dominance and economic conditions will not change due to recent fluctuations. 3. No need to switch just for the sake of switching, continue to adhere to the dominance of economic conditions In terms of allocation, there is no need to switch just for the sake of switching; the relative strength of economic conditions and the relative change in performance remain the core clues. Whether it is from the changes in the unanimous profit expectations after the first quarter, the latest industrial enterprise profit data, or surveys of institutional investors, the consensus on economic conditions remains highly concentrated. AI computing hardware sectors (optical communications, semiconductors, PCBs) still have the strongest consensus, with some advanced manufacturing sectors (AI upstream equipment, battery energy storage, ships) seeing improved economic expectations. In addition, upstream resource products (many of which are driven by AI demand, such as minor metals, energy metals, chemicals, etc.) have also shown growth. Internally within the AI sector, from the perspectives of congestion, leading price comparisons, and overseas mapping, the recent downward trend in the North American computing chain (optical modules, PCBs) due to overseas influences, combined with economic advantages, makes the July earnings verification period more strategically cost-effective. Domestic AI software services (IT services, network security, enterprise software) have further room for upward spread compared to overseas companies. In addition, the recent rise in US pharmaceutical stocks (innovative drugs, AI medicine, brain-to-machine interfaces) as a low-level sector that needs catching up, requires attention when mapping to A-shares. Risk warning Volatility in economic data, softer-than-expected policy easing, less rate cuts by the Federal Reserve, escalation of geopolitical tensions, etc. Source: WeChat public account "XYSTRAREGT", Author: Xingzheng Strategy Team, GMTEight Editor: Chen Qiuda