Fundamental Thinking featured image

TSMC Leads Emerging-Markets Returns due to AI Enthusiasm

Momentum strategies were a powerful force for global equities in 2024, primarily ones that followed artificial intelligence (AI) and its extended value chain. The trend was particularly acute in emerging markets (EMs).

For the calendar year, virtually all of the returns in the MSCI EM Index could be traced to just five stocks, led by TSMC, which on its own contributed nearly half of the index’s return. The dominance of these stocks was most evident in the fourth quarter. Every sector except Information Technology fell, and the only regions that rose were the Middle East, which eked out a small gain, and Taiwan, home to a number of prominent tech stocks including TSMC (the “T” stands for Taiwan.)

TSMC rocketed on surging demand for AI-related chips and a seemingly unassailable leadership position in the industry following failed attempts by competitors to take market share in advanced process technologies. The company has more than 80% market share in leading-edge semiconductors globally and fabricates almost all of the chips designed by NVIDIA. TSMC expects AI-related revenue to grow at a cumulative rate of 50% over the next five years.

Hon Hai Precision, the giant Taiwanese electronics contract manufacturer, also saw its shares soar as growth prospects have been boosted by the demand for AI servers. The other top-five contributors, however, came from China and outside the AI-momentum trade: Tencent, Meituan, and Xiaomi.

OOOM featured image

Salesforce, ServiceNow Benefit from Next Phase of the AI Wave

By now, it is clear to most investors that the companies benefiting from the move to artificial intelligence (AI) include more than just a couple of chipmakers. One analogy for the increasing size and breadth of the AI industry is a tsunami: As the wave of corporate investment in AI builds, and the underlying hardware, foundational machine-learning models, and early-stage software applications all continue to improve, a broader set of tech providers has been able to benefit from the associated demand.

For example, Alphabet and Meta are among large cloud-services companies—also known as “hyperscalers”—that are building both the physical infrastructure and large-language models that are needed for AI technologies to be adopted by corporations more broadly. While the servers in their data centers continue to require powerful graphics processing units designed by NVIDIA and manufactured by TSMC, these hyperscalers are also designing their own custom chips, called application-specific integrated circuits (ASICs), which they are developing in partnership with chipmaker Broadcom, boosting that company’s growth outlook as well.

Advances in AI technology are also benefiting enterprise-software providers, as the rise of agentic AI in the fourth quarter brought the long-term promise of computers with human-like problem-solving capabilities into sharper focus. Agentic AI is capable of sophisticated reasoning and can automatically come up with ways to solve complex, multi-step problems—a significant step forward by models such as OpenAI’s o3 and Google’s Gemini 2 as compared to the more limited tasks performed by the previous generation of AI technology. Therefore, agentic AI is likely to be more broadly useful in business software, with companies such as Salesforce and ServiceNow well positioned to offer powerful tools based on this technology. At a recent event in San Francisco, Salesforce Chief Executive Officer Mark Benioff said the company is focused on its agentic-AI platform, Agentforce, adding, “The only thing we’re going to do at Salesforce is Agentforce.”

OOOM featured image

Chipmaking Is Getting More Complex. Daifuku’s Smart Monorails Keep Fabs Running Smoothly

In semiconductor manufacturing, a single speck of dust poses a threat to production. It’s why cleanrooms, the sterile labs where silicon wafers get etched and cut into pieces, and then packaged as finished chips—with thousands of steps in between—contain few humans. To reduce the risk of contamination and defects, materials are largely transported by automated monorail systems that travel along the ceiling.

Source: Daifuku.
While advances in generative artificial intelligence (AI) have put a spotlight on the companies that design and manufacture chips, as well as their data-center customers, providers of cleanroom technology play an increasingly critical role in a world of high-performance computing. Not only is the industry for cleanroom automation characterized by an attractive competitive structure, but new trends and challenges in chipmaking are also improving the growth outlook for this specialized material-handling technology. One player in particular may stand to benefit, and that is Daifuku.

OOOM featured image

Chips of the Trade: TSMC, Samsung Benefit from AI Demand

In international markets, a big theme of investor interest relates to companies developing the underlying technology that powers AI. This includes the designers and manufacturers of the advanced semiconductors necessary to run AI, as well as producers of semiconductor manufacturing equipment and providers of the critical computing infrastructure required by AI systems.

Expectations are that semiconductor industry revenue growth will accelerate to annualized double-digit levels this decade, spurred by demand for AI chips. This would be a growth rate well above levels that we’ve seen since the mid-1990s, with predictions that the roughly US$50 billion dollars of AI chips sold in 2023 could rise to US$400 billion dollars of sales before the end of the decade.