Pro Invest

Semiconductors: The AI beneficiary

Artificial Intelligence, or AI, is a global fixation – one that has driven the Nasdaq to outperform every other major world equity index this year.

AI is having its moment, and a knock-on effect of its popularity is that semiconductor companies have reemerged in the spotlight as one of AI’s crucial initial beneficiaries.

Semiconductors are the elements that conduct electricity voltages between that of typical conductors and insulators. These electrification properties enable semiconductors to be used in computers and other electronics to control the flow of electricity, store memory and convert energy.

This isn’t the first time semiconductors have been popularised recently. Pandemic supply constraints and lifestyle adjustments like working from home drove demand in the early years of the COVID-19 pandemic, but there was a brief period of quiet where they weren’t performing as well.

The factors that caused this initial demand have not gone away, but the demand for AI is eclipsing any other driver of semiconductor use.

So, the Nasdaq is benefiting from this upward swing, but why exactly are semiconductor companies benefiting so much from the AI boom?

Chips – Logic and memory:

To understand why semiconductors are benefiting from AI’s popularity,
it’s important to understand the role the materials have played in the advancement of technology. Computer chips are made from semiconductors and are used for calculating and storing data.

In order for AI to function, it too relies on the use of chips. The two major types of chips are logic and memory.

Logic chips, sometimes referred to as microprocessors, are the ‘brains’ of electronic devices. In this context, logic implies ‘processing’, and the chips perform instructions and tasks that the device needs to execute. Every electronic device needs a storage unit, and memory chips serve this purpose.

The process of manufacturing semiconductor chips is one of the most complex engineering and scientific feats that humans have achieved. To the untrained eye, this process seems like pure magic.

The process begins with the design of the chip, where engineers use complex software to design the circuit of transistors. Post design completion, the blueprint is sent to a foundry (a metal castings factory) for manufacturing. The ‘magic’ of semiconductors lies in their ability to scale and continually get smaller over time.

In 1965, businessman and engineer Gordon Moore proposed that the number of components on a semiconductor would double every year.

In 1975 he revised his forecast that the rate would double every two years. Since then, his prediction held, and has widely become known as Moore’s law. Some manufacturers have made their chips smaller and more efficient than others. In the world of chips, the leading-edge is a term used to describe the smallest and fastest chips produced.

Companies with the highest profit margins are the leaders of the leading-edge space, as they have access to large capital investment that in turn funds research and development costs. As the number and type of devices worldwide seems to exponentially grow, so does demand for these chips. Data centre servers, smartphones and other advanced computing devices account for a large amount of this increased demand.

AI is the newest driver of this demand.

Chip hungry Artificial Intelligence:

Artificial Intelligence models including OpenAI’s ChatGPT require vast amounts of computation, memory, and data to run.

Having always experienced high demand, leading-edge chips regularly eclipse analyst estimates. Leading-edge graphics processing units (GPUs) are used to train the world’s most advanced AI models, and as AI has developed at a significant speed, so has the demand for GPUs. The newest generation of chips are more economical than previous generations, meaning training these models is more efficient per dollar spent.

For AI to operate effectively it needs to run off chips that are efficient and economically viable. For investors looking for semiconductor-related investments they can look towards companies that are tied to leading-edge semiconductors.

This includes Nvidia and AMD in chip design, ASML and KLA Corp who provide machines that measure chips, and TSMC which manufactures chips as a contract manufacturer. Predictions for the future of the semiconductor industry are bright, with McKinsey and Company assessing a trillion-dollar valuation for the industry by 2030.

Of all the recent winners that have emerged from the AI boom – tech giants, cloud titans, even cyber-criminals – chipmakers may just be the biggest.

Tobias Taylor is Director, Wealth Management Adviser at Jarden.

The information and commentary in this article are provided for general information purposes only. It reflects views and research available at the time of publication, using external sources, systems and other data and information we believe to be accurate, complete and reliable at the time of preparation. We make no representation or warranty as to the accuracy, correctness and completeness of that information, and will not be liable or responsible for any error or omission. It is not to be relied upon as a basis for making any investment decision. Please seek specific investment advice before making any investment decision or taking any action. Jarden Securities Limited is an NZX Firm. A financial advice provider disclosure statement is available free of charge at