The author of this article notes that while there remains considerable focus on AI and the chip industry, artificial intelligence-related investment hasn't been a sure-fire bet, and that shrewd stock selection and positioning is crucial.

The following article by Kwai San Wong, analyst, global equities at Sarasin & Partners, explores the investment implications of AI and, specifically, the impact on semiconductors. 

The editors of this news service are pleased to share this content, and invite readers who want to enter the conversation to jump in! Email tom.burroughes@wealthbriefing.com and amanda.cheesley@clearviewpublishing.com

Semiconductors, driven by AI’s need for power-hungry graphics processing units (GPUs), have seen significant growth, but challenges remain. 

Semiconductors may not be the first things that come to mind when we think of our tech devices, but they are critical to our modern technologies and are present everywhere. In short, they are the fundamental components – used to convert and amplify electronic signals – that make the chips which power our digital economy today.

Semiconductors are found in our everyday devices, from smartphones, to computers, appliances, cars, planes, and industrial equipment. Every day, we create products that have more and more semiconductor content. It’s no surprise, therefore, that global semiconductor sales have been growing at 5 to 7 per cent annually in the past 20 years – they are big business and investors are keen to gain exposure.

Nvidia and the rise of AI
The semiconductor industry has dominated the investment headlines in 2024, and much of the focus has been on one stock in particular, Nvidia. From its formative years in creating graphics cards for gaming and multimedia, the company has grown tremendously to hit a multi-trillion-dollar market capitalisation. Today its (GPUs) are used extensively to power the development of artificial intelligence.

The pace of AI development has surpassed even the most optimistic predictions since the launch of ChatGPT in late 2022. This was a key event in kick-starting a new investment cycle involving big technology companies which currently rely on Nvidia’s chips to build AI models. Nvidia and its suppliers have benefited significantly from this trend – the PHLX Semiconductor index (SOX) has climbed by approximately 25 per cent year-to-date (to 30 September), beating the 20 per cent performance from the S&P 500 in the same period.

The semiconductor industry is known for its pronounced cyclicality; the high-tech foundries needed for production are very expensive and time-consuming to build, but once established can produce huge volumes of computer chips around the clock. 

This high ratio of upfront to operating costs means that relatively small swings in demand, due to the ebbs and flows of the economic cycle, are often amplified into huge swings in profitability for the semiconductor industry. Understandably, the growth of AI has sparked a record-breaking uplift in profits for the industry leader, Nvidia.

In 2025, consensus is expecting Nvidia to earn $160 billion from sales to the datacentre market, which is almost 10 times higher than in 2022. This spectacular growth has made it more difficult for investors to forecast the next inflection point in the industry, and so every quarterly result is now closely watched by investors.

While individual investment cases differ, a commonality among all of them is the idea that AI will spark the next cycle of investment in semiconductors, particularly as tech companies need to build the hardware before they can make more progress on applications. This trend was the same in the previous technological cycle (e.g. personal computers, internet, mobile and cloud computing).

How are semiconductor companies embracing sustainability?
GPUs are power hungry, and so a huge drain on energy resources. In order to provide more computing power for AI training and inferencing, Nvidia has increased the power consumption with each new generation of its products. This, in turn, has meant significant incremental demand for power generation from the data centres which house these GPUs.

While the semiconductor industry has a long track record of making more efficient chips with each new generation, the sheer amount of power needed by these AI data centres is definitely something it needs to find a sustainable long-term solution for. This is a key area of concern for the biggest users of semiconductors. 

For example, Microsoft recently signed a 20-year deal with Constellation Energy to re-open previously decommissioned nuclear reactors at Three Mile Island to supply power to its AI data centres. Apple meanwhile has shared details of new reusable data-centre air filters; the company has a stated its aim of being 100 per cent carbon neutral across its products and supply chain by 2030.

What are the risks of investing in an AI future?
Despite the rapid progress in AI development in the past two years, we still have not seen a clear path for all related companies to turn into profitable businesses. Naturally, return on investment is front of mind of every investor.

While we have seen promising use cases for AI in areas such as customer service agents and code generation, helping software developers to be more productive, the question remains whether the scale of investment can be justified.

The market will certainly watch the launch of upcoming models such as GPT-5, Llama 4 and Gemini 2 closely, and analyse how much more capable they can be. If these new models fail to deliver an increase in abilities commensurate with the outlay on development and operational costs, then sentiment towards AI investments could cool. However, if the 'scaling laws' hold and these models meet expectations, we expect AI to unleash the next era of computing, with transformative effects on society and the global economy.

We have kept a close eye on the recent volatility in the share prices of semiconductor companies, and concerns about a potential ‘bubble’ in all things related to AI. However, as it stands today, we are happy to hold a basket of stocks that we see as long-term investments that remain well placed to profit from ongoing technological progress.

The future for semiconductors
Semiconductors are becoming more and more vital to our modern technologies. We expect semiconductors’ importance to increase in the future and we will continue to attempt to identify the right investments in the sector for our clients.

We remain cautiously optimistic about developments in AI, though we are keeping a close eye on the profitability of related companies, and valuations within the wider tech sector.