A research report indicates that the recent sharp downturn has altered the landscape in terms of stock prices, valuations, and holdings, but has not changed the divergence in fundamental performance. For high-quality hard tech assets that have now become attractively priced, this represents a favorable entry point for long-term allocation. Furthermore, crowding in most core AI sectors has receded to historical lows, with these stocks experiencing significant overselling relative to their overseas counterparts over the past month. Within the market, increasing attention should be paid to leaders in the North American computing power supply chain, such as optical communication firms.
The past month has witnessed the most severe global market volatility since the start of the current AI bull market. As the world searched for reasons and de-levered, caught in a negative feedback loop of liquidity and panic, the market was forced to trade time for space, awaiting clearer signs of a sentiment bottom and more definitive guidance from fundamentals. This week marks a crucial turning point. Global stock markets have simultaneously initiated self-rescue measures to curb the liquidity feedback loop, and several key catalysts have materialized, offering a reason for the market to return to rational analysis. Therefore, the choice between AI and non-AI sectors ultimately hinges on a return to the core of sector comparison: fundamentals, stock price positioning, valuations, and holdings. Clarifying what has changed and what has not after this downturn will provide a clearer path for future allocation.
What Has Changed After the Downturn: Stock Prices, Valuations, and Holdings
The most significant change after the steep decline is the convergence of the previous divergence between AI and non-AI sectors. The substantial accumulated gains and overly concentrated holdings were the primary triggers for the global AI correction. However, the correction has now rebalanced the structure and digested holdings. In terms of stock prices, valuations, and holdings, the market is no longer characterized by divergence. From a stock price perspective, the high-growth index has turned negative year-to-date, underperforming the dividend yield index. The dispersion in year-to-date performance among first-level sectors has dropped to the fifth-lowest level since 2010. On valuations, the 2027 price-to-earnings (PE) ratios for many typical tech and consumer leaders have fallen to the 10x to 20x range, making them no longer expensive from a long-term perspective. Regarding holdings, short-term market sentiment crowding indicators have clearly shifted, with previously low-level consumer and dividend sectors now showing elevated crowding, while most tech and growth sectors have seen crowding drop to historical lows. Following the correction, position pressure in AI-related sectors has likely been substantially digested. For active mutual funds, AI-related sectors have generally experienced a 25% to 50% pullback since July, suggesting a significant reduction in allocation through both price declines and active portfolio adjustments. Margin debt in the electronics and communication sectors has fallen 23% and 26% from recent highs, respectively, exceeding the de-leveraging magnitude seen in previous rounds since 2024, indicating that high-risk capital pressures have been largely alleviated. As a result, the divergence between AI and non-AI in terms of stock prices, valuations, and holdings has been considerably narrowed. With these dimensions now flattened, the core focus for assessing allocation returns to the advantage of fundamentals.
What Has Not Changed After the Downturn: The Fundamental Divergence
The fundamental divergence between AI and non-AI sectors has remained unchanged. The current AI adjustment is primarily driven by a negative liquidity feedback loop amplifying volatility and panic, with no substantial change in fundamentals. The market now needs clearer information to rebuild consensus. With the recent release of earnings reports from North American cloud companies and the conclusions of the Politburo meeting, the market should have a clearer judgment on the fundamentals of both AI and non-AI sectors. For AI, the feared scenarios of disappointing earnings and a slowdown in AI capital expenditure have not materialized. The latest earnings reports from North American cloud firms paint a picture of AI returns significantly exceeding expectations, a growing order backlog intensifying supply-demand imbalances, and long-term capital expenditure visibility. Even after substantial upward revisions since the second quarter, the proportion and magnitude of earnings beats for U.S. tech stocks remain at historically high levels. Capital expenditure guidance for 2026 continues to be revised upward by hyperscalers, with the four major cloud service providers' combined 2026 capital expenditure estimated at $720 billion to $745 billion, implying a growth rate of around 80%. While no quantitative guidance was given for 2027, none signaled a reduction. The returns on AI investments have exceeded market expectations, and the backlog of orders for cloud services is growing rapidly, exacerbating supply shortages and ensuring the sustainability of AI capital expenditure. For non-AI sectors, the focus is on domestic policy support in the second half of the year. The latest Politburo meeting conveyed a policy stance that is more proactive on the margin, but prioritizing the implementation and effective use of existing policies. The mention of "counter-cyclical adjustment" and "timely planning of incremental policies" is more positive than before. If fundamental downside pressure increases in the third or fourth quarter, incremental policies may be deployed to provide support. However, the necessity for significant additional fiscal or monetary easing appears limited. Therefore, the fundamental divergence between AI and non-AI is likely to persist in the second half of the year, but the recent sell-off has brought their valuations back to similar levels. A simple PE-G comparison shows that while the 2027 PE ratios of many typical tech and consumer leaders are in the 10x to 20x range, they correspond to expected growth rates of over 30% and around 10%, respectively. Once the market calms, it will reassess the allocation value of tech leaders based on this.
Current Allocation Strategy
The sharp downturn has altered the landscape in terms of stock prices, valuations, and holdings, but has not changed the fundamental divergence. Therefore, for high-quality hard tech assets that have become attractively priced, this represents a favorable entry point for long-term allocation. Crowding in most core AI sectors has receded to historical lows, and these stocks have experienced significant overselling relative to their overseas counterparts over the past month. Within the market, increasing attention should be paid to leaders in the North American computing power supply chain, such as optical communication firms. Additionally, based on marginal changes in fundamentals and recent catalysts, current recommendations focus on the AI mid-to-downstream sectors, which are seeing narrative shifts and increasing catalysts; oversold AI upstream materials, including new materials, small metals, energy metals, fiberglass, and plastics; advanced manufacturing sectors such as shipbuilding, battery storage, innovative drugs, and power grids; and cyclical stocks with alpha, including chemicals, textile manufacturing, non-bank financials, industrial metals, and beer, where earnings expectations have been revised upward since July.
Comments