Published in partnership with Pictet Asset Management
The artificial intelligence revolution is still in its infancy, but already there is scarcely a business in existence not touched by digitalisation and the application of technology in its operations, in one form or another. What we’ve seen so far is just the tip of the A-iceberg.
AI’s implications are far-reaching: from producers of the rare earths needed to manufacture semiconductors, which then go into every data centre; to hyperscalers that deliver cloud services to businesses and consumers; to software application developers; and to businesses that use those applications to better deliver goods and services to consumers.
As an investment strategy, an investible universe that incorporates all traditional businesses trying to exploit technology, including AI, to modernise or streamline their operations is simply too broad.
Hundreds of thousands, if not millions, of businesses “could be seen as ‘digital companies’, if you take the loosest definition”, says Pictet Asset Management’s (Pictet AM) senior investment manager, active thematic global equities, John Gladwyn.
Conversely, a number of focused investment strategies have been developed to cash in on the development of AI as either a revolutionary (or at least, a very rapidly evolutionary) technology that will change the world in ways we can’t even guess at, let alone invest in with certainty.
Somewhere in between is an approach that recognises AI as an accelerant to the already healthy growth of digital technologies. Gladwyn says a distinction must be made between companies and businesses adopting technology to streamline existing processes to drive efficiency gains, and businesses that have technology, including AI, at their core.
“Because we invest in…digital-first companies, that leaves us with an investment universe of about 250 companies,” Gladwyn says.
For example, he says, “Walmart was primarily a supermarket operating company, and then it has adopted the walmart.com website and [become an] e-commerce business because it had to, because that’s where the market went”.
“Amazon.com started as an e-commerce platform,” he says. “And that’s where we draw the line: between those that are at their core digital businesses, versus the pre-existing, let’s say legacy businesses.”
CalPERS deputy chief investment officer, capital markets and operations, Dan Bienvenue says the fund considers technology and AI “through lots of lenses”.
Like any large organisation that deals in making sense out of large volumes of information and data, CalPERS has to consider the impact of AI on its own business. As an investor, CalPERS invests in AI-linked opportunities directly – the chip makers, the cloud providers, the software developers and so on. Some of this it does itself, and some it does through external managers.
The derivative opportunities
Bienvenue says there are also “derivative” opportunities – businesses that will use AI to streamline their processes and systems and, in some cases, to reimagine their businesses operations from top to bottom.
“Similar to the rise of cloud years ago, the rise of the Internet years before that, like home computing – all of that stuff – I actually think that the bigger implications are going to the indirect ones, the derivative ones,” Bienvenue says.
“It’s not so much the AI itself, it’s the way that AI leverages how they do their business. There will be a number of disruptors and disrupteds, and that’s actually arguably as big or bigger a theme in terms of the implications for our portfolio.”
Bienvenue says there will even be second, third, fourth and fifth-derivative implications of AI “exactly the way that there have been for past big technological steps forward”.
“I would put AI in a similar category, and specifically generative AI is another technological step function forward. That’s going to have lots of second and third and fifth order derivatives as a result.”
AI investment opportunities are a subset of the broader technology universe, but Border to Coast Pensions head of equities Will Ballard says “equally, the benefits of AI can reach further than the technology universe itself”.
“The technology sector can cover everything from AI [to] software and services, to hardware manufacturers and semiconductor equipment,” he says.
Ballard says “every company is different and our role as investors is to be able to distinguish between them, understand their competitive advantage, what they could be worth, and determine whether they are an attractive investment or not”.
“Just like there is an opportunity within the automotive supply chain, from the suppliers to the manufacturers and then the distributors, all the way through to the taxi operators or logistics providers, understanding their position within their business landscape, their bargaining power with their suppliers, the barriers to entry to other competitors, and the demand from their customers is essential,” he says.
Chipmaker Nvidia, for example, is “in an exceptionally attractive position, being the sole supplier of leading-edge hardware necessary for the ongoing AI revolution”, Ballard says.
Meanwhile, however, Google is in competition with Amazon and Microsoft and others in cloud services.
A traditional business that might use AI to improve its own operations or to disrupt its competitors must be understood in the context of its own particular industry, Ballard says.
“Just like we must understand the sustainability of Nvidia’s dominance, so we must understand the competitive advantage and persistence an early adopter of AI might have within a more traditional setting,” he says.
Catering to business, catering to consumers
Pictet AM’s Gladwyn says that on top of identifying technology-native businesses, it also organises potential investment targets into two further groups: those set up to cater to business customers, such as Microsoft; and those set up to cater to consumers, such as Amazon.
“And then we have the bottom layer of enabling technologies, which would be more semiconductors, for instance, that are critical enablers on which everything else depends,” he says.
Gladwyn says the trends that AI serves to accelerate are already established.
“The big trends for the strategy have been cloud migration – movement of workloads into the cloud – and digital transformation,” he says. “AI is an accelerant to both.”
“It’s one of the reasons why this investment cycle has been so far benign – what we have seen in our universe is that the winners have just remained winners.”
It’s a strategy that deliberately ignores the potential gains that might come from companies that do successfully transform themselves.
“For every one that does it, there are many more that do not,” Gladwyn says.
“You can spend so much time micro-analysing 10 companies that claim they’re on a transformational journey, but ROI on that time is just not as good as focusing on technology-centric companies.”
Gladwyn says Pictet AM classifies technology-centric companies into four broad categories, in a kind of conceptual stack.
“You have software at the top; you then have the large language model [LLM], which is kind of the intelligence layer,” he says.
“Below that you have the cloud infrastructure, which itself is built on hardware, including of course Nvidia’s GPUs.”
Gladwyn says the attraction for investors is that all of those layers will change over time, but the timing and pace of that change will vary, layer by layer.
“Because things change, it gives you the opportunity to try and get ahead of that,” he says.
“But what’s so interesting to us is that there is a big difference between when the different layers of the technology stack experience change. Before you run the railways, you have to build the railroads. We’re now building the railroads. Nvidia is benefiting before a software company is benefiting, because you can’t have one without the other.
“The value-added part is trying to differentiate which areas benefit and when.”
This is the “impossible question that we are all grappling with”, says Border to Coast’s Ballard.
“The first assumption is that data-heavy sectors such as insurance or finance might be quick to benefit,” he says.
“What we are seeing is that there are signs that the impact is much wider than that. Just like with the advent of the internet, the scale and impact on our lives of AI is going to be tremendous. It is likely that it touches everything we do, there is no sector that is not going to be impacted.”
Ballard says even sectors considered unlikely to benefit, such as mining which might seem tied more closely to commodity prices, “could experience changes to how they go about everything from the geological surveys at the start of a project all the way through to the way they sell and distribute the final processed product”.
Pictet AM’s digital strategy generally holds between 35 and 50 stocks, and currently the investment portfolio is at the lower end of that range.
Gladwyn says that as a public-equity investor Pictet AM’ s digital strategy is “quite excited” about the current cycle. Big, publicly listed companies like Google and Microsoft are likely to be the biggest beneficiaries “because they have the talent, the data, the infrastructure and the money to actually take advantage of this technology”.
“This cycle is fairly different from previous cycles, in which the beneficiaries have been the smaller, more nimble private companies,” he says.
“The characteristics of the companies we invest in are growth businesses. They will have above average top-line growth. These companies are also highly profitable. In the past the earnings per share (EPS) growth of this group of companies has outperformed quite nicely the top line growth of either the MSCI World or MSCI IT Index.
“We believe this combination of structural growth and strong financials is highly attractive to end investors.”
The lessons of history
Academic and researcher Professor Ajay Agrawal says figuring out a potential path for the take-up, impact and investment opportunities of technology and AI might be easier if it can be compared to earlier technological revolutions, such as the invention of electricity. Agrawal says both AI and electricity are “general purpose technologies”.
As University of Toronto Rotman School of Management Geoffrey Taber Chair in Entrepreneurship and Innovation, Agrawal says the take-up of electricity by manufacturers was initially low; after roughly 20 years after electricity was invented, only around 3 per cent of factories were using it and when they were if was for marginal gains, such as replacing gas lamps with electric lamps.
Part of the reason was the sheer weight of investment that had been made in traditional processes, including steam engines.
Agrawal says that gradually, more entrepreneurial factory owners, and those building new factories from the ground up, began to adopt electricity in preference to steam. It revolutionised the design and construction of factories. No longer were the massive columns needed that supported the driveshafts of the massive steam engines and took up factory-floor space. Now, every machine in a factory could have its own power supply and motor, vastly reducing the amount of time lost when one steam engine driving multiple machines broke down.
Changes in design and construction in turn revolutionised manufacturing processes and streamlined production. Before long, factories powered by steam were no longer economically viable.
Applying a technology like AI to a company’s existing processes, but leaving the processes essentially unaltered, could be quick and generate profit gains in order of 20 per cent, Agrawal says. Designing processes and systems from scratch, and putting the new technology at the core, takes longer but the gains could be in the order of 500, 600 or even 700 per cent.
Gold mines and rabbit holes
Ballard says the key to keep on top of the fast-changing opportunities that a breaking technology wave like AI might deliver is to “stay informed and be open-minded”.
“Our fund managers and analysts are always talking to CEOs, industry experts, academics across all different sectors try to get a better understanding of what trends they are seeing and what they are thinking of doing,” he says.
“The main point we stress is that we recognise that we don’t know all the answers, so we approach these big, transformative questions with an open and eager mind. We recognise that there is no such thing as certainty, and so informed, critical, probabilistic thinking is crucial to good decision making.”
CalPERS’ Dan Bienvenue says, to use a baseball analogy, AI is in about the third inning. It will develop quickly, often in unexpected directions. Investors will find themselves going down rabbit holes as some developments play out productively and some do not. Bienvenue says that’s inevitable.
“In order to find the gold mines, you have to go down some of the rabbit holes,” he says.
“Ex-ante, those can’t be differentiated. You try to differentiate and you work with smart people that can try to differentiate, but the only way to avoid every rabbit hole is to avoid the gold mines also.”
Bienvenue says in addition to company-specific or industry-wide impact of new technology on portfolios and investment returns, there will be broad economic implications that will touch all areas of an investor’s portfolio.
“Again, I’m the equity guy, but I think they are also exciting,” he says.
“There will be both a multiplier effect of just the money that’s being spent on AI – and there’s a lot of money being spent – and some of that is being spent in unproductive places; we can just all acknowledge that not every dollar invested will be productive.
“There’s not only that implication that I think is probably a per cent or two [added] to GDP, but then there’s this whole big productivity side [and] we haven’t even seen all the places that it’s going to manifest itself. That also is going to drive economic growth.
“Through the lens of the investment outcomes, that becomes self-reinforcing. I would call it a virtuous cycle. I’m the eternal optimist, and I’m very optimistic about these impacts on a go-forward basis.”
