Guosen: Commodities benefit from AI-driven resource pricing power reconstruction, maintaining optimism for the medium to long term.

Guosen released a research report stating that the current global asset allocation logic is undergoing a profound shift, with the macro narrative transitioning from valuation repair to profit realization. Equity assets are being driven by expanding US stock profit margins, the storage cycle of South Korean stocks, and the significant advantage in the bond-stock ratio of A-shares, providing strong upward elasticity. Commodity assets are benefiting from AI-driven resource pricing restructuring and are seen favorably in the medium to long term; the bond market is constrained by the fiscal premium of US bonds and the rising risk appetite of Chinese bonds, weakening defensive measures in the long end. Guosen's main points are as follows: Commodities: Transition from real estate and infrastructure to computing power and electricity, restructuring the pricing logic of resource products Currently, global commodities are showing strong upward momentum, with a fundamental shift in logic. Traditional manufacturing sector pricing indicators such as copper and silver are deviating from PMI trends, reflecting a global economy transitioning from old dynamics to new-quality productivity. The AI industry follows a new path of "expanding computing power - increasing electricity demand - resource consumption", with large-scale hardware construction providing rigid support for conductive and thermal materials. In addition, the expectation of trade protection triggering physical hoarding demand and the "security premium" brought about by geopolitics give resource products a currency risk-hedging attribute. In the long term, demand elasticity has systematically increased, and despite short-term pressures for excessive gains, the structural pattern of commodities remains solid. Bond market: Chinese bond long end sees volatile repair, US bonds focus on fiscal risks In terms of Chinese bonds, the slowing growth momentum of the macroeconomy supports the bond market, but the ultra-long bond yield spread is at historically low levels and needs to adjust valuations. Accompanied by the stronger equity market driving increased risk appetite, coupled with the release of pressure from ultra-long bond issuances in January, the momentum of long-term varieties is under short-term pressure and shows a pattern of volatile repair. As for US bonds, macroeconomic data shows divergence, with the resilience of the service industry delaying interest rate cuts. President Trump's direction to purchase $200 billion in MBS as a QE-like measure substantially raises long-term inflation expectations, focusing on the fiscal dominance risk. Strategically, it is recommended to control long-term risk exposure, adopt a "short duration + steepening" strategy, and hedge against sticky inflation using TIPS. A-shares: Policy guidance shifts "crazy bull" to "slow bull", profit momentum begins to spread from point to face Measures such as raising margin ratios to reduce leverage clearly signal the regulatory shift to a "slow bull", and short-term corrections do not change the medium to long-term upward trend. The comparison of major asset classes shows that the A-share dividend yield relative to bond yields has been at an extreme high of 7% since 2005, making it extremely attractive. The bull market is entering its second phase, with improvements in fundamentals shifting from point-like repairs represented by the STAR 50 index to sector-wide diffusion. With the implementation of anti-internal competition policies and systematic interest rate cuts, it is expected that by 2026, the net profit attributable to the parent of the entire A-share market (excluding financial sector) will significantly increase by 10% year-on-year, shifting market logic from mere valuation repair to performance and valuation driven by dual wheels. US stocks: Realization of AI efficiency dividend, expansion of profit margins solidify bull market foundation Guidance for the 4Q25 earnings season of US stocks is optimistic, with media forecasts indicating a potential 20% growth in earnings for the quarter, significantly higher than market consensus. The key factor is the substantial realization period for AI technology applications in reducing costs and increasing efficiency. Giants like Amazon and IBM are optimizing cost structures through AI, driving a stepped-up increase in operating profit margins. Although valuation premiums still exist, the optimization of profit quality through improved enterprise efficiency strengthens the foundation of the bull market. As profit momentum spreads from the "tech seven giants" to secondary technology and energy sectors, US stocks demonstrate strong resilience. Future market trends will depend more on the continuous expansion of profit margins rather than simple inflation of valuation multiples. Japanese and South Korean stock markets: Structural transformation of earnings in Japan, South Korea enters a super cycle of memory chips As important pillars of the Asian technology chain, both Japanese and South Korean stock markets have seen significant upward revisions in earnings prospects. The Japanese stock market benefits from governance reforms and benign inflation cycles, with profit margins of large companies expected to reach a historical high of 9.4%, providing a safety net even though the forward PE ratio of the Nikkei 225 is close to a high of 24 times. The South Korean stock market directly benefits from the super cycle of AI memory chips, with profit expectations rising by nearly 30% since the second half of 2025, leading emerging markets. The explosive increase in foreign capital allocation to the South Korean semiconductor sector reflects a global reallocation within the AI hardware industry chain, with South Korean stocks becoming a new growth pole for Asian technology momentum. Risk warning Uncertainty in the pace and extent of overseas monetary policies; geopolitical and trade disruptions; unexpected tightening of liquidity; and retreat of technical valuation bubbles.