As AI giants continue to act as a cash cow for investors, asset managers have started to search for — if not greener — cheaper pastures that are benefitting from the boom.
Investing upstream and downstream from AI companies, in businesses that are either putting the technology into practice or solving demand needs to advance AI, has caught attention from fund managers as the industries experience waves of growth to keep pace with the likes of Nvidia.
Chris Rossbach, CIO of J Stern & Co, said while his World Stars Global Equity fund continues to invest in Nvidia, he is also being mindful of which sectors will benefit most from AI down the road, and how he can build out the industrial sector of the business.
“Over time, companies that adopt AI will see benefits in terms of productivity, efficiency gains and new product development in ways that would otherwise have been unachievable,” Rossbach said.
“That is why we see great opportunities today to invest in the ‘picks and shovels’ companies that facilitate the development of AI and to identify companies that will benefit from AI use cases in their business models.”
The ‘picks and shovels’ mindset has drawn Rossbach particularly towards social platforms, where AI can enhance targeting specific audiences for companies like Meta, as well as the healthcare industry. Rossbach said AI could create cost saving of 25% to 50% for drugs in developmental stages.
See also: Amundi, Norges Bank and Union Investment call for an end to arms exports to Israel
“AI can help in all stages of the drug development process. It can help identify drug targets faster, it can help simulate and predict which compounds are best for lead discovery and optimisation, it can help with candidate selection to choose molecules and measure the interaction, and it can help test the safety and effectiveness of a drug,” Rossbach said.
Felix Wintle, manager of the VT Tyndall North American fund, has turned his attention not just to businesses that he thinks will benefit from the growth, but companies where capital expenditure from big tech names are already giving a leg up.
“We actually have an underweight in tech relative to the benchmark. But most of the tech stocks we own are, surprise surprise, related to AI,” Wintle said.
“One of the interesting stories around AI more broadly is that there’s an awful lot of CapEx happening today to make AI possible. This CapEx is quite easy to understand, because we essentially need much bigger data centres to handle all the demands.”
Because of the energy intensity of AI data centres, water cooling systems are necessary rather than being cooled by AC units. However, this system comes at a high expense, accounting for near 40% of the cost of a new data centre, Wintle said. Yet the systems are necessary for advancement, causing Magnificent 7 companies such as Microsoft, Google, Amazon and Meta to pay for them.
The result is a projected $200bn in capital expenditure on data centres for 2024, up 35% from a year before. One of the companies benefitting from the building frenzy is Comfort Systems, an industrial stock Wintle purchased 15 months ago.
“The CEO told me Google is just writing them cheques, and that they are building as fast as they possibly can,” Wintle said.
See also: FCA remodels capital raising rules in bid to boost UK markets
In the past year to 26 July, the stock has grown by 81.37%. Alphabet has gained 28.6% in the same period.
Rossbach has also keyed into getting a piece of the AI pie through the industrial sector, emphasising the growth that AI requires from the sector even compared to technology from just a few years ago. While a data centre five years ago may have required 20-30MW of power, a large AI data centre today clocks in at near 100MW.
“The industrial sector has multiple ways to benefit from AI. First, as an enabler to companies across many sectors from consumer to utilities to manufacturing, industrial stocks can benefit from the development of AI. They can expedite product design, optimise production processes, manage complex global supply chains and enhance predictive maintenance techniques,” Rossbach said.
“Secondly, industrial companies are benefiting from the resurgence in physical asset spending, after decades of underinvestment, in both the public and private sectors. And in their own businesses, industrials are driving greater efficiency, productivity and profitability.”
The demand has led to particularly strong growth in two of Rossbach’s holdings. Eaton, an intelligent power management solutions company has grown 46.8% in the past year and Amphenol, a connector and sensor company, has grown 44.6%.
Michael Crawford, CIO of Chawton Global Investors, said that while its difficult to predict what businesses AI will benefit in the future, he believes a few monopolies have developed within the semiconductor supply chain. Because chips have shrunk, the equipment used to create them has become increasingly complex and has narrowed the industry down to fewer and fewer companies, Crawford said.
“For example, the mirrors used in Dutch company ASML’s lithography machines have to be so smooth that if they were the size of a football pitch, the greatest imperfection would be the size of a pea. Only Carl Zeiss, an associate company and supplier can produce these. Lithography is central to the manufacturing process and has been the key driver of Moore’s law,” Crawford said.
“As a result of these factors, the semiconductor supply chain has evolved so that generally only one or two companies globally dominate the equipment and process control technologies. They are in effect monopolies and oligopolies. ASML is a monopoly; there is simply no other company that can compete with their latest EUV machines.”
ASML, which creates chip-making equipment, has grown its share price by 23.7% in the last year, while KLA Corporation, which provides measurement equipment for quality control, is up 71.6%.
“While it is difficult to predict how exactly artificial intelligence will develop and which companies will prove successful in deploying it, the semiconductor equipment supply value chain, that underpins it, is set in stone. It has monopolistic characteristics and is seeing strong growth in demand as GPU-related infrastructure starts to replace CPU,” Crawford said.
“We believe investing in these companies is the best way to prosper in the revolution in computing power. What’s more; it is only just getting started.”
