The market is now transitioning from a period of exceptional volatility to a more sustainable phase, requiring investors to moderate their expectations for returns. We categorize the performance landscape into three key thematic lines: the North American AI supply chain, exports to non-US markets, and domestic demand-driven sectors.
AI price chain rebound nearing completion, patience needed for new narratives
After a significant two-week recovery, the risk-reward ratio for the AI price chain has notably diminished. The median forward P/E for the domestic AI price chain sample has risen from 1.18 times the cycle's starting point in late July to 1.5 times by August 14, reducing the expected recovery-to-downside ratio from 6 times to just 1.2 times. While the underlying industry trends and pricing logic remain intact, the presence of both trapped floating losses and recent speculative gains makes a double-top scenario and a one-time valuation re-rating unlikely. The truly worthwhile opportunity lies in identifying new highs, not in re-gambling on cyclical peaks.
Recent market narratives around AI for Science and RSI have generated some buzz, but these are likely insufficient to drive a broad sector revaluation. Historical patterns show a significant lag between technological breakthroughs and capital market rallies. Like the delay between the launch of reasoning model o1 and the subsequent rally in mid-2025, investors should wait for concrete product penetration and verified market validation before committing capital. The risk of missing a major rally is low when consensus is high but technology is immature. While new catalysts like Grok 4.6's performance may support a rebound in core AI stocks, they are unlikely to lift the entire sector without a non-linear leap in model capability.
Overseas export risks persist, especially from EU trade disputes and FX losses
Trade tensions between China and the EU are expected to remain unresolved in the near term, with September-October being a key period for potential progress. Measures such as anti-subsidy duties on EVs and medical device procurement restrictions have already been implemented. The EU's Carbon Border Adjustment Mechanism (CBAM) and other regulations will add compliance costs across the entire business chain, from supply chain due diligence to local production requirements. These costs, while not immediately causing a sharp drop in exports, will persistently compress profit margins and capital returns, creating a long-term valuation overhang for companies with significant European exposure.
As the mid-year reporting season approaches, investors must closely monitor the impact of RMB appreciation on exchange losses. Data from the State Administration of Foreign Exchange shows net forex settlement by banks reached $320.5 billion in the first half of the year, 1.37 times the full-year 2025 level. For 384 non-financial firms that have disclosed their half-year reports, the financial expense ratio in Q2 2026 increased by 0.47 percentage points year-on-year, with companies having over 30% overseas revenue seeing a much larger rise of 1.04 to 1.74 percentage points, primarily due to exchange losses. This is a critical risk for the export-driven performance theme, which remains a key driver for A-share earnings.
Domestic AI hardware shines, but sentiment is tied to the global AI chain
Within the pure domestic demand category, domestic AI hardware (like semiconductor equipment and materials) stands out as a rare bright spot, with cumulative gains of +22.8% and +1.3% since June, significantly outperforming the Philadelphia Semiconductor Index (-3.2%) and KOSPI150 (-19.1%). However, this sector's pricing is becoming relatively full. Following the recent rebound, some funds have rotated from the North American AI chain into domestic AI hardware, but positioning congestion remains high, which may limit future valuation upside. The sector's performance is not entirely independent of the global AI narrative.
Uptrend intact, but the pace of gains will slow
The market's volatility spiked significantly in Q2 2026, with the number of stocks doubling in price reaching 275—the highest since Q1 2007. This is a decadal anomaly, not a normal state. While the industry trends for both the US-facing and non-US export chains remain upward, factors constraining valuation expansion are increasing. It is unlikely to see a systemic, large-scale valuation uplift in the near term. Going forward, market opportunities will likely be more about valuation repair for stocks with solid earnings support. The process of market de-volatilization will also be a process of optimizing the shareholding structure and re-pricing high-quality assets, laying a foundation for a more sustainable medium-to-long-term rally. Investors must now return to a normal expectation of reasonable returns, moving away from the illusion of doubling gains.
In terms of portfolio allocation, within the technology sector, we recommend using the rebound in AI price-chain stocks to rotate into core assets like gas turbines, wafer manufacturing platforms, and semiconductor equipment, prioritizing "volume certainty" over "price volatility." For non-technology sectors, the focus should be on increasing exposure to chemicals, metals, innovative pharmaceuticals, and leading brokerages with overseas expansion potential.
Key risks
Key risks include escalating US-China friction in technology, trade, and finance; weaker-than-expected domestic policy effects or economic recovery; a faster-than-expected tightening of global liquidity; escalation of regional conflicts in Ukraine and the Middle East; and a slower-than-expected digestion of China's real estate inventory.
Comments