Beyond the Hype: Understanding the AI goldrush
As artificial intelligence continues to advance, its impact on companies and economies worldwide is becoming increasingly profound. This paradigm shift presents both opportunities and challenges for investors.
To help shed light on this important area, we examined the impact and implications for investors. There can be no hotter topic for long-term investment than artificial intelligence (AI). Yet the phrase itself is broad and can be rather vague.
Let’s strip it back.
It was 1955 when John McCarthy, an American computer scientist, coined the term “artificial intelligence” at the Dartmouth Summer Research Project, he defined AI as “the science and engineering of making intelligent machines”.
Since then, we’ve had to become a little more specific. AI, or the simulation of human intelligence processes by computer systems, has now evolved to include learning, reasoning, and self-correction.
AI has a number of subfields, all with a variety of useful applications. Perhaps the most pervasive are machine learning and natural language processing (NLP) which may not be familiar terms, but are, in all likelihood, already affecting our daily lives.
The latter will be familiar to users of ChatGPT, although most of us are exposed to NLP via more subtle means. Machine learning is sometimes likened to the training of animals. Namely, computer programmes that learn through trial and error, often with more speed, rigour and accuracy than a human could ever achieve, particularly on specific repetitive tasks.
What’s the role of AI in today’s world?
The objective of AI is to create systems that can efficiently perform tasks that would typically require human intelligence, such as visual perception, speech recognition, decision-making, and language translation, among others.
AI technologies have wide-ranging applications across industries. In fact, it’s hard to think of an industry that won’t already be using AI to some extent.
If we take banking and insurance companies for example, they already employ a range of machine learning processes to assess the risks within their insurance or lending business. They can be helpful for fraud detection, or credit default forecasting. While natural language models have the potential to improve businesses’ interactions with customers.
We’ve probably all had infuriating experiences dealing with chatbots, but that might not be for much longer. AI has given firms high hopes about their ability to provide a better customer experience by quickly getting customers the answers they need. This functionality will be welcomed by service industries such as banking and insurance, but it’ll also be widely employed by firms needing to handle routine queries from customers.
With compelling applications like these likely to become a reality across industries, it begs the question, how can investors reap the benefits?
Investing in AI
Intuitively, it seems likely that AI will make investors money. Taking the example of customer service chatbots, if they can do the work of humans faster and more accurately, the ongoing cost of running the computer will probably be less than the staff currently dealing with queries. Call centre staff can then focus their time on resolving more complex or unusual issues.
So, over time, the initial investment in AI will be exceeded by reduced costs.
Or will it?
To answer that question, we need to consider how competitive the industry is and how replicable the AI strategy is.
If the industry is competitive and all competitors have a similar ability to employ AI-enabled customer services, then they’ll likely compete for business. They’ll attempt to gain market share by cutting prices until the gains from lower costs have been effectively passed on to their customers. Of course, lower prices for customers are very welcome, but it doesn’t mark out a great investment strategy.
What are the most profitable ways to gain from AI?
Put simply, a company that has a strong competitive position will generally be better placed to benefit from lower costs driven by efficiency initiatives. Others may even be able to use AI to generate new sales, rather than simply savings costs.
The best-known users of AI are technology-enabled companies such as Netflix, which provides customised recommendations based upon users’ viewing history, searching history, ratings and demographic characteristics. But there are now several online content streaming platforms, and the success of one or another seems likely to depend upon the appeal of their content libraries rather than their use of AI to access it.
Booking.com is employing a similar approach and envisions AI playing a central role in creating a more personalised and easier travel planning experience. This includes features like its AI Trip Planner, which is expected to be enhanced over time based on traveller interactions.
For many technology-enabled companies, the secret to their competitive advantage is not the AI itself, but the data the AI gets to interact with.
AI produces economies of scale that were much more muted in previous stages of economic development. Now, if you have a lot of customers, you have a large potential pool of customer interactions with which to train your AI model. Within the right business model, AI can produce natural monopolies, which accumulate, analyse and leverage data that would otherwise be costly or impossible to collect.
Customer interactions are an example of data. They’re a helpful byproduct of doing business which, if recorded and analysed, can provide insights about what works, for whom, when and why. And that’s AI’s bread and butter.
There are some companies that have a distinct data advantage. Google for example, offers search for free, but builds up useful data on search preferences, which it can use to sell targeted advertisements. AI is used to determine what your search history suggests about products and services you might be partial to. It then uses AI to determine the value to advertisers of your attention, based upon how likely their advertisement is to succeed.
What Google knows about your search history, Amazon knows about your shopping preferences. Both these companies amass data and analyse it, and their competitive advantage lies mostly in their access to enormous pools of high-quality data, which come through their dominant positions in internet search and online retail, although they are very sophisticated analysts too.
By contrast, social media platforms accumulate huge amounts of data on their users but the commercial value of it might be harder to gauge and, as the data pool swells, the cost of fishing out the useful conclusions rises too.
Companies with limited access to such data can still determine a niche in which they too have a competitive advantage. RELX amasses data for use by professional and business customers in fields like science, medicine, law and engineering. It curates high-quality datasets and offers the tools to analyse them. Rather than focussing on scale as Amazon and Google would, RELX focuses on credibility and reliability to reflect the potential use cases of their data.
How to navigate the AI gold rush
It’s clear that AI will become increasingly common in almost all industries. And how it’s implemented will be key to a firm’s success.
Failure to implement AI effectively could put a firm at a disadvantage, whereas implementing it immaculately may have the opposite effect. But for most companies, that advantage will dissipate unless there’s something to stop peers from imitating their success.
However, identifying how effectively AI is employed within a business isn’t the only way to prosper from it financially.
Investors seeing a secular change in the way the economy works may try and pick the potential winners, or they may focus on the indirect beneficiaries – an analogy used in investing called “picks and shovels”.
Picks and shovels is a reference to the U.S. gold rush of 1848-1855, when many tried and almost as many failed to strike it rich looking for gold. Gold prospectors had limited success and endured extreme hardship in what was a highly competitive search. Ultimately, too many prospectors caused competition that diluted gains and increased costs.
The parties who did gain were the wholesalers and merchants (such as Levi Strauss), who supplied prospectors with the tools they needed, at prices that reflected their short supply.
Picks and semiconductors
Today, the picks and shovels of the AI boom are semiconductors. If AI
makes computers intelligent, then their reliance on semiconductors is analogous to our reliance on brain matter.
Semiconductors are used in all modern digital machines, not just intelligent ones. So, as we find digital functionality creeping into an increasing array of goods, the growing demand for semiconductors becomes structurally significant. That being said, despite this structural uptrend, there will likely be cycles of oversupply according to a cycle of design, development and production.
Although there are several markets for different kinds of semiconductors, their production peaked in 2022. Since then, much of the industry has been suffering a downturn. However, things now seem poised to recover, benefitting from the increasing structural demand for different semiconductors well suited to the deployment of AI.
It’s this anticipation that has propelled companies like Nvidia, who design and manufacture graphic processing units (the most powerful semiconductors required for powering AI), into the highest echelons of the global stock market. Nvidia is just one part of a supply chain that stretches from chip design through to manufacture.
Integral to this supply chain is ASML in the Netherlands, which builds extreme ultraviolet lithography machines. These machines enable the precise etching of miniaturised integrated circuits that drive more advanced and efficient digital processes.
The future of AI integration
AI is an enormous subject with amazing potential and some risks. Companies cannot afford to ignore it, as being left behind could be costly.
But good implementation of AI alone isn’t the key to a successful business strategy or investment. Rather, for most companies, the question is what can AI do to provide an enduring advantage over competitors or, in the case of equipment providers, can they be beneficiaries of the expanding use of AI?
It is these questions and more being discussed by our investment partners. If you would like to discuss this or your investments further – please don’t hesitate to contact us.