A recent series of monthly reports from leading actively managed private equity funds, including Fushisheng, Hunjin, Dongfang Harbor, Chongji, Juming, Yuanlesheng, Gaoyi, and Renqiao, has been released during the interim reporting season.

Performance data reveals that institutions heavily invested in artificial intelligence (AI) are leading the pack, while those focused on domestic demand and consumer sectors are lagging significantly. AI remains the dominant market narrative, yet significant internal disagreements exist among these institutions. Broadly, consensus is narrowing while divergence is widening; however, most institutions acknowledge that beyond AI, it is currently difficult to identify a second industry direction capable of absorbing large-scale capital flows.

Dongfang Harbor, a veteran multi-billion-dollar private fund established in 2004, has achieved a year-to-date return of 25%. Adhering to an investment philosophy of "growing with great enterprises," its founder Dan Bin has consistently maintained heavy AI positions throughout this market cycle, a strategy validated by performance.

In his latest monthly report, Dan Bin dissected the industrial logic of AI, providing a key data point: the combined capital expenditure of North America's four major tech giants is projected to reach approximately $730 billion by 2026, nearly doubling year-over-year—following a 67% increase already projected for 2025. In his view, this is not mere speculative hype but substantial, real investment flowing into AI infrastructure.

However, Dan Bin also offered a contrarian perspective in the report. He expressed particular concern about the erosion of profit margins for downstream Cloud Service Providers (CSPs) due to rising storage costs—the proportion of CSP hardware capital expenditure allocated to storage has surged from around ten percent at the onset of the AI wave to over 50%. Concurrently, the prosperity of AI applications heavily relies on the rapid decline of token costs. "These two trends are moving in opposite directions," he noted.

In other words, Dan Bin identifies a structural tension within the AI supply chain: rising storage costs upstream are squeezing the profit margins of downstream cloud service providers, while the flourishing downstream AI application ecosystem requires continuously falling costs. This represents a critical contradiction that requires ongoing monitoring.

It is noteworthy that Dan Bin previously pointed out on social media that the current U.S. stock market trend is highly illustrative. While the once tightly-grouped "Magnificent Seven" tech giants underwent a collective correction, market capital did not exit; instead, it executed a massive rotation into high-growth sectors such as semiconductor equipment, advanced manufacturing processes, and computing power hardware. The capital's stance is clear: actively embracing hardcore technology with tangible industrial iteration and earnings realization, resulting in a classic structurally divergent market.

The A-share market is now fully replicating this K-shaped logic:

The upward stroke of the "K" – Hard technology sectors like semiconductors, AI computing power, and import substitution are becoming core holdings for institutions and northbound capital, supported by global AI capital expenditure expansion, policy backing, and sustained order fulfillment.

The downward stroke of the "K" – Traditional sectors lacking technological moats, suffering from weak domestic demand and overcapacity, and devoid of growth narratives are experiencing continuous capital outflows, valuation downgrades, prolonged weak volatility, and poor investment experiences.

Dan Bin stated that this bifurcated "ice and fire" market dynamic is not a short-term sentiment fluctuation but a generational repricing driven by the iterative upgrade of the silicon-based economy replacing the traditional carbon-based economy. Global capital markets have moved past broad-based rallies and declines, formally entering an era of "the strong getting stronger, the weak getting weaker."

Amid this extreme K-shaped divergence, market timing becomes less critical. The core of investment lies in selecting the right sectors and positioning correctly within them. Adhering to the high-growth hard technology trend while avoiding weak traditional assets is the path to aligning with long-term industrial trends and capturing the enduring opportunities within this market split.