29 July 2026
Ryan Davies, CFA Senior Portfolio Manager
Michael P. Evans, CFA Managing Director, Equity Client Portfolio Manager


Semiconductors belong to one of the most specialised yet globally integrated industry chains. From design, equipment, and materials to manufacturing and commercialisation, the production of a smartphone chip alone spans many countries across continents, creating tremendous opportunities for companies, consumers, and investors. With semiconductors increasingly becoming the backbone of an artificial intelligence (AI) race few are prepared for, understanding this sector is key to unlocking where the next wave of technology competition is heading.
Many of us are familiar with how chatbots like ChatGPT and DeepSeek work. Beyond these simple uses, however, few are aware of how AI is already being embedded into everyday life. Some examples of early-stage enterprise AI use cases include:
These examples demonstrate that AI adoption is still in its infancy. The runway for growth looks potentially long, and the opportunities that lie ahead are substantial.
The historical demand drivers of semiconductors have shifted over time: Personal computers (PCs) dominated in the 1990s, followed by handsets in the 2000s. Today, AI is emerging as a pivotal driver of future growth of semiconductors (represented by computing and data storage in the next figure), alongside automotive, consumer, and industrial demand. Valued at nearly US$800 billion in 2024, the semiconductor industry is now projected to reach US$1.6 trillion by 2030.
![]()
In recent years, we have seen a surge in spending by AI model firms and hyperscalers (i.e., large cloud service providers running data centre networks), which has led to an obvious concern: is there risk of a bubble? We don't believe so; AI infrastructure investment remains foundational rather than late-cycle.
AI spending is concentrated among a handful of hyperscalers, but they are investing on behalf of billions of consumers. Given this combination of concentrated spending and broad-based beneficiaries, we think the current spending cycle is more sustainable than the historically more cyclical semiconductor booms.
In our view, the sustainability of AI expenditure and the growing range of AI use cases provide tailwinds across the semiconductors and AI value chain. AI and data centres lead the pack, with strength expected to continue into 2027 as a multi-year buildout plays out. Clusters of graphics processing units (GPUs) are driving near-term demand across data centre connectivity, compute, and semiconductor equipment broadly. The outlook looks similarly promising for Industrials and Internet of Things (IoT), while other markets remain more selective.
Looking ahead, humanoid robots are emerging as a longer-term opportunity, with decades of manufacturing and product data now being used to train machines capable of performing tasks autonomously. Meanwhile, power management is fast becoming a critical constraint, as AI's energy appetite increasingly outpaces available grid capacity.
![]()
Semiconductors are historically volatile. Although annual corrections of 10-15% are typical, we see these as buying opportunities. Despite the AI narrative, we believe current valuations remain in line with historical averages1, while the earnings of the PHLX Semiconductor Sector Index (SOX) and the index itself have moved closely in tandem over the past 20 years.
The pattern is evident: semiconductor content has been expanding into many areas of daily life and may remain a long-running trend, with AI potentially acting as an accelerator.![]()
1 Source: Bloomberg, as of 15 June 2026. Semiconductors represented by Philadelphia Semiconductor Index.
AI innovation: Asia helps build many of the world’s technologies behind it
When electricity first arrived, the world built the necessary infrastructure – power plants, transmission lines – before the real transformation could take hold. A similar process is happening with artificial intelligence (AI). Today's massive investment in chips, data centres, and power grids is laying the foundation for a potential expansion in AI application that could take years to develop. In our view, the discussion is increasingly shifting from whether AI adoption will continue to how the enabling infrastructure is being built. Asia appears to be playing an important role in that development.
China Fixed Income: From deflation to reflation: what comes next?
Not another bubble: How semiconductors are powering a real future
Semiconductors sit behind almost every modern experience – from smartphones and cars to cloud computing and today’s AI tools – yet they remain largely invisible to most people. They are more than chips only, and the demand is being supported by several long-term forces. We believe that today’s semiconductor excitement is not a repeat of the dot-com bubble, as investment is tied to real infrastructure and revenue-generating services. And the opportunity is broader than a handful of headline AI names.
AI innovation: Asia helps build many of the world’s technologies behind it
When electricity first arrived, the world built the necessary infrastructure – power plants, transmission lines – before the real transformation could take hold. A similar process is happening with artificial intelligence (AI). Today's massive investment in chips, data centres, and power grids is laying the foundation for a potential expansion in AI application that could take years to develop. In our view, the discussion is increasingly shifting from whether AI adoption will continue to how the enabling infrastructure is being built. Asia appears to be playing an important role in that development.
China Fixed Income: From deflation to reflation: what comes next?
Not another bubble: How semiconductors are powering a real future
Semiconductors sit behind almost every modern experience – from smartphones and cars to cloud computing and today’s AI tools – yet they remain largely invisible to most people. They are more than chips only, and the demand is being supported by several long-term forces. We believe that today’s semiconductor excitement is not a repeat of the dot-com bubble, as investment is tied to real infrastructure and revenue-generating services. And the opportunity is broader than a handful of headline AI names.