AI boom spurs new opportunities; navigating supply chain shifts is key

Ongoing AI evolution spurs new opportunities, albeit volatility is inevitable
Growing adoption of AI agents drives CPU market; demand for chips and memory remains robust
Asian tech, traditional sectors also benefit from AI boom; Asian tech valuations about 40% below the US
To identify evolving opportunities from the AI shift, active management may be more effective than a passive strategy

AI continues to evolve, moving beyond simple computational tasks towards complex, multi-step enterprise applications. While technological advancement inevitably brings market volatility, it also creates new investment opportunities. Notably, the beneficiaries of the AI boom extend beyond US technology companies to include many Asian technology firms, as well as traditional sectors that benefit from rising AI-related capital expenditure.

In just over a year, AI applications have shifted from generative tools to systems capable of supporting complex workflows, commonly referred to as 'AI Agents'. During the early stages of generative AI development, large language models perform tasks such as translation and content generation by relying heavily on graphics processing units (GPUs), which are capable of processing vast amounts of data at high speed. As technology matures, AI agents start gaining popularity. Guided by the objectives set by humans, these systems can autonomously search for information, consolidate data, coordinate multiple steps and make decisions. Such applications require not only the computational power of GPUs, but also central processing units (CPUs) to manage data processing and system coordination, leading to a more integrated GPU-CPU computing architecture. Amid growing corporate adoption of AI agents, the market expects the global server CPU market to at least double by 2030 compared with last year.

AI data centre capital expenditure continues to rise, driving the development of new technologies such as co-packaged optics (CPO). Compared with traditional copper wiring, CPO uses optics interconnects to transmit data at even higher speeds while reducing power consumption and signal attenuation, thereby improving operational efficiency and system stability. Since the beginning of this year, the share prices of two US optic networking equipment manufacturers have surged by 86% and 120%, respectively.

Growing AI investment has also strengthened the existing supply chain. Buoyed by mounting demand for AI servers, NAND and DRAM prices have rose by over 200% in the first four months of the year. Globally, these two products are primarily supplied by two key South Korean manufacturers, reflecting their strong pricing power. As with the semiconductor supply chain, an upstream US chip design giant and a midstream Taiwanese major chipmaker have both demonstrated sustained operational strength and share price gains. Less in the spotlight is the outsourced semiconductor assembly and test (OSAT) segment, which has also benefited from the AI capex cycle. Notably, a Taiwanese leader nearly doubled its firstquarter earnings from a year ago.

Apart from technology shares, the proliferation of AI infrastructure is also driving many traditional industries, such as power and utilities, cooling systems, industrial equipment, and network infrastructure. While these sectors may not be directly involved in AI application development, they play a key role in supporting AI computing, data transmission, and data centre operations. The growth potential of these sectors has yet to be fully recognised by the market.

AI investment opportunities extend beyond the US, with many firms based in Asia. Some of the Asian technology companies have seen extraordinary share price gains of 140% to 988% over the past year. Yet, for many, earnings growth has outpaced share price performance, leaving valuations across Asia's tech sector about 40% lower on average than those in the US, suggesting valuations remain attractive.

To capture AI's structural growth potential, a passive investment strategy may not effectively capture the shifting opportunities that emerge as technology evolves. Active management, supported by a disciplined stock-selection framework, could help uncover investment opportunities with earnings growth prospects and reasonable valuations. Additionally, AI-themed assets tend to be more volatile. In light of concerns that a single-asset allocation could carry higher downside risk, a multi-asset allocation strategy can help manage short-term market volatility, enhancing portfolio resilience and its ability to absorb market shocks.