Harvest Fund's Fang Han: In the Year of the Horse, optimistic about AI diffusion, supply and demand improvement, and the main theme of pro-cyclical recovery

【Editor’s Note】

2026 marks the beginning of the “14th Five-Year Plan” period, as China’s economy enters a new development stage.

Under the new circumstances, the voices of foreign investment banks bullish on China are continuous. Goldman Sachs recommends overweight positions in A-shares and Hong Kong stocks in 2026; JPMorgan Chase has upgraded the ratings of mainland China and Hong Kong markets to “Overweight”; UBS believes that policy support, improved corporate profits, and capital inflows may drive up A-share valuations. These judgments reflect international capital’s recognition of China’s economic transformation direction and development prospects in 2026, and also suggest that as winter turns to spring, global capital may flow toward the East.

The “Chief Connection” market outlook for 2026, titled “Spring Water Flows East,” also conveys this meaning. In the outlook, the “Chief Connection” studio will interview dozens of authoritative economists, fund managers, and analysts to discuss their views on China’s economy in the new year and analyze new investment opportunities.

“From the expectations of institutions at the end of 2025, this is a year with relatively strong consensus in the past three years,” said Fang Han, Director of Stock Strategy Research at Harvest Fund, in an exclusive interview with The Paper.

According to Fang Han’s analysis, there are two clear major consensus points in the 2026 market: first, market form—the continuation of a structural market trend; second, the main structural theme—the greatest consensus still revolves around the AI technological revolution. However, behind this consensus, he also pointed out three major core divergences that the current market needs to pay attention to.

Regarding industry allocation in 2026, Fang Han focuses on three main lines: AI diffusion, supply-demand improvement, and pro-cyclical recovery.

The core logic driving market operation remains unchanged

“Whether it’s the ‘core assets’ rally back then or the ‘track investment’ craze now, fundamentally, they both reflect that the Beta characteristics in the A-share market structure are becoming increasingly prominent,” Fang Han stated clearly. He analyzed that, against the macro background of China’s shift to medium-high speed growth, the difficulty of achieving high global cycle growth has increased, and structural changes in industrial transformation have become the most important feature at each stage of economic growth. This makes capturing phased industrial trends a necessary answer for high returns in each mid-term market cycle.

From “core assets,” “dual carbon revolution,” “localization,” to the current “AI revolution,” “revaluation of strategic resource commodities,” and “leading enterprises going overseas,” the change of terminology has not altered the essence of the pricing logic. Fang Han emphasized that, in a historical context where the stock market shows more structural rather than global features, the core logic driving market operation has always been closely linked to “industry trend opportunities that align with the era background and structural transformation directions.”

However, he also observed that with the highly developed social media, widespread ETF coverage of niche sectors, and the rapid dissemination of information, capital feedback efficiency and consensus formation speed have far surpassed previous market cycles.

“We know that, while efficient pricing is achieved, in a market environment where Beta characteristics are increasingly evident, it can also lead to excessive pricing of certain market segments driven by group sentiment, followed by correction,” Fang Han said.

Two major market consensus points and three core divergences

Fang Han frankly stated that, based on recent institutional expectations, the current market has formed a relatively high degree of consensus in recent years. In the context of two consecutive years of positive performance and relatively clear structural features, market thinking inertia tends to produce consistent expectations.

“The reason why consensus can form is mainly because it usually has strong self-consistency and is widely recognized by the market,” Fang Han explained.

According to his analysis, the two major consensus points for the 2026 market are quite clear: First, market form—the continuation of a structural market trend. This judgment is based on a relatively positive internal and external policy environment, the potential for residents’ savings to enter the market, and more companies crossing profit cycle inflection points; the judgment of structural features stems from the market’s own operating rules after two years of valuation repair, as well as the overall recognition of high-quality development in the capital market, which discourages rapid and excessive leverage.

Second, the main structural theme—the ongoing focus on the AI technological revolution, from infrastructure investment surges to application breakthroughs across the industry chain; the second major consensus is on commodities benefiting from diversified AI demand, geopolitical supply vulnerabilities, and the loosening of the US dollar credit system.

However, the more solid the consensus, the easier it is to overlook potential pitfalls. Fang Han pointed out three major divergences in the current market:

First, whether valuation expansion “will not last more than three times” and whether the “high allocation curse” around the tech sector will be broken during the historic AI technological revolution.

Second, whether rising prices of commodities, and possibly spreading to “low-level” energy prices like oil, will break the Federal Reserve’s path of rate cuts throughout the year.

Third, whether the re-pricing of hundreds of trillions of residents’ deposits maturing within the banking system, and residents’ reallocation choices, will lead to massive shifts of savings into stock assets.

Three main lines for layout: AI diffusion, supply-demand improvement, and pro-cyclical recovery

The market is simultaneously paying attention to emerging growth driven by new productive forces and the transformation value of traditional industries. For 2026 industry allocation, Fang Han focuses on three key areas.

First, the diffusion effect of new technological changes. This year’s global AI investment will shift more from AI input to applications that generate benefits (revenue or efficiency). He especially emphasizes that epic-level AI infrastructure investments are creating revaluation opportunities for traditional industries: benefiting from supply-demand gaps in storage and semiconductor equipment; benefiting from North American power shortages in power grid equipment, machinery, new energy, and materials; and segments like liquid cooling, which began to realize profits in 2026.

Second, industries with stable demand and reduced supply pressure, such as lithium batteries, military industry, offshore wind, dairy and flavorings, and aerospace.

Third, high-odds pro-cyclical assets. As the economy surpasses pressure peaks, sectors closely related to macroeconomic prosperity—such as real estate, food and beverages, and discretionary consumption—are expected to gradually recover their fundamentals with economic rebound, increasing their win rate.

Three potential risks to watch closely

Under a “rationally optimistic” tone, Fang Han’s analysis of potential risks is particularly sharp.

First, the contradiction between market behavior and policy guidance. Fang Han believes that the A-share market itself tends to accelerate its upward rhythm, with retail investors and short-term funds chasing gains, combined with related public opinion influences, creating a prominent conflict with the pursuit of stable, high-quality development advocated by the capital market.

Second, whether the AI main theme will face changing expectations. He warns that North American cloud service providers’ relentless capital expenditure might, under the backdrop of overdrawn corporate cash flows, reach a turning point in marginal improvements of large models. If doubts about sustained CAPEX in North America grow, the valuation and sentiment of the AI chain in A-shares could face pressure.

Third, external macroeconomic reversals. If the US economy smoothly lands softly and enters an expansion cycle amid a new round of fiscal stimulus, 2026 could mark the end of the current Fed rate-cut cycle. Such external changes could suppress global liquidity and pro-cyclical assets in A-shares.

(Source: The Paper)