A Review of 2025 and Outlook for 2026

Investment Review

In 2025, major global capital markets performed well overall, with the S&P 500, Hang Seng Index and CSI 300 rising by 16.4%, 27.8% and 17.7% respectively. The AI-related industry chain contributed significantly to structural alpha.

However, the investment process was not as simple and smooth as the results suggest. At the beginning of the year, the market was concerned about an oversupply of AI computing power due to the efficiency improvements in models driven by Deepseek; in April, the black swan event of tariffs impacted the overall market; in the second half of the year, AI investment opportunities shifted from "comprehensive diffusion" to "structural differentiation," with sub-sectors such as the Google ecosystem, storage, and networking performing exceptionally well, while returns on purely computing power-related assets were relatively flat; entering November, discussions about an "AI bubble" intensified significantly. Looking back at the whole year, if the judgments made at these key junctures had been flawed, the final portfolio returns would have been significantly different.

In terms of specific investment strategies, at the beginning of the year, facing concerns about AI computing power oversupply triggered by Deepseek, while recognizing the increase in equivalent computing power supply, we focused more on its role in accelerating the penetration of inference applications. We also judged that as AI applications evolve from simple dialogues to more complex tasks, the overall computing power demand will continue to grow exponentially. Therefore, we maintained a high allocation to AI-related assets.

In early April, after the tariff shock triggered a rapid market decline, we judged that the situation was more likely a "controllable shock" than a systemic crisis. With risk premiums significantly rising and the market nearing the lower end of historically manageable decline ranges, we chose to maintain risk exposure and conduct structural adjustments rather than adopt a defensive contraction strategy.

Entering the second half of the year, as AI investment opportunities shifted towards structural differentiation, based on our assessment of changes and certainties in the industry chain landscape, we increased our allocation to areas such as the Google ecosystem, storage, and networking, while relatively reducing our reliance on a single computing power expansion logic.

Overall, we maintained a relatively high position throughout 2025, focusing on the internet, semiconductors, cloud computing, consumer electronics, and fintech. AI spans multiple sectors and remains a key area of ​​focus for us, with investments concentrated on leading AI application companies with long-term competitive advantages, key infrastructure, and leading chip manufacturing companies.

Opportunity Outlook for 2026

Macroeconomics and Valuation: Focusing More on Fundamental Drivers

Looking ahead to 2026, the macroeconomic environment is expected to provide weaker support for asset valuations than in 2025. While interest rate cuts remain the medium-term trend, factors such as high government leverage, dollar credit constraints, and sticky inflation will likely lead to a more restrained and limited pace of rate reductions. This means that both the scope and speed of a long-term decline in risk-free interest rates are limited.

Against this backdrop, current market risk premiums are already at historically low levels, limiting the potential for further valuation increases to drive overall returns. Investment opportunities in 2026 are more likely to stem from improvements in corporate fundamentals, evolution of business models, and strengthening of structural competitive advantages, rather than systemic valuation expansion.

AI：While the overall volume remains uncertain, structural opportunities still abound.

The market's discussion about a "bubble" surrounding AI investment is somewhat justified. Current expectations of approximately $3 trillion in AI capital expenditure over the next three years are likely to fall short of expectations in terms of funding pace, data center delivery capacity, and the speed at which the industry chain can absorb the investment. Therefore, simply betting that the total amount of AI infrastructure investment will be no less than or significantly exceed expectations carries low odds and a low probability of success.

However, at the same time, another clear trend deserves more attention: the continuous improvement of AI capabilities and the rapid decline in unit task costs are driving the exponential expansion of AI applications. Even if the cumulative capital expenditure in the future is lower than the most aggressive expectations, such as reaching $2 trillion, its absolute size is still enormous and has profound significance for structural opportunities.

Historically, this evolutionary path is not unfamiliar. Take Apple as an example: the iPhone was released in 2007, and its annual sales reached approximately 200 million units in 2015 before gradually peaking. Afterward, the growth in both sales volume and unit price tended to level off. However, even after smartphone sales peaked, Apple's market capitalization still increased approximately tenfold. This wasn't simply due to inflated valuations, but rather the continuous evolution of its business structure: from a single hardware company to a multi-product, multi-service brand and ecosystem, significantly strengthening its competitive advantage.

During this process, besides the company itself, its supply chain system also fostered numerous structural opportunities with potential for tenfold growth after sales peaked.

The development of the AI ​​industry may exhibit similar characteristics. Even as the overall growth potential of infrastructure gradually converges, structural opportunities will persist in the long term, distributed across multiple levels and sub-sectors, as the industry evolves from computing power, storage, and networks to applications and ecosystems.

Other key areas: Finding differentiated sources of Alpha

In addition to AI, we will continue to focus on fintech, consumer and emerging markets as important sources of supplementary Alpha.

• Fintech and Financial Innovation: Finance is a large-scale, high-return industry. Any efficiency improvements, product innovations, or business model transformations can potentially create opportunities for substantial long-term returns.

• Consumer Sector: Since 2025, some consumer sectors have been impacted by the expected impact of AI on middle-class employment, leading to valuation compression even for companies with sound fundamentals. Market concerns are not unfounded, but after a general decline in valuations, significant differentiation may emerge, providing new entry points for long-term investment.

• Emerging Markets: Differences in growth structures, policy space, and industrial stages among different economies provide diversified sources of alpha for global investment. In the future, we will more actively explore structural opportunities with high-return characteristics.