With the significant appreciation in the value of artificial intelligence (AI) stocks like Nvidia and Advanced Micro Devices, as well as cloud players including Microsoft, there have been a slew of very fair questions from investors around whether it is too late to invest in Loftus Peak portfolios.
The first point to note is we do not hold the highest weights in stocks that have the greatest expected appreciation in revenue or earnings. Those higher-weighted positions are reserved for companies with more modest expected revenue growth, but where in our view that revenue growth has the greatest certainty. In practical terms, what this means is that our investment in AI has been staged to catch returns from distinct parts of the AI cycle. The details of how we determine these parts can be read here. Presently, the portfolios are weighted towards the tools and enablers of AI and in time we expect to raise exposure to beneficiaries of AI. An investment in Loftus Peak’s portfolios is very different from an investment in Nvidia or AMD. This investment approach has stood the test of time over the past ten years.
There is hype, but there is also opportunity
Commentators sometimes throw around the term Artificial General Intelligence (AGI), the idea of AI having broad skills analogous to those of a human – especially the ability to learn on the fly. This is compounded by the various performative robots like those showcased by both Tesla and OpenAI (in partnership with robot company Figure). In our view, these are at the more speculative end of the market; they are investments with greater uncertainty and so deserve a lower allocation of capital, meaning smaller weights in Loftus Peak Portfolios.
For every egg-grasping robot there is a necessary value chain of wafers, cables, processing power, sensors, memory and cloud compute. Often the valuations these players command may not seem related to their end market. The AI fervour is such that even some (but not all) companies along this value chain are being priced for the occurrence of fringe optionality.
Super Micro Computer (not held in the portfolios we manage) is an example. The company assembles the physical servers from Nvidia chips before they arrive in a datacentre. By our reckoning it is trading ahead of Nvidia, which is nevertheless its primary revenue driver. In a little over two months, the company rose +318.3% before cooling off to a +259.9% year-to-date figure (as of the 27th of March).
Although Nvidia is a holding in the portfolios we manage, we have reduced the weighting as it is currently richly valued. We covered this in more detail in last month’s insight. The weight to Nvidia in Loftus Peak portfolios was reduced in favour of Taiwan Semiconductor Manufacturing (TSMC) and Qualcomm; companies that have exposure to AI and cyclical recovery, but are not trading at the same valuations.
The same framework that led us to reap significant returns from Nvidia and AMD draws us to positions in TSMC and Qualcomm. Respectively, these companies supply the wafers and handset chips to enable further development of AI as well as enabling the port of AI models onto the edge. Regardless of whether the more unproven end of AI interest comes to fruition, the likes of TSMC and Qualcomm will underpin the AI disruption that is already underway.
What this means for portfolio construction
Of course, that still means that Loftus Peak has a relatively high exposure to AI, when taken at a whole-of-portfolio level. This is our intention. We believe that AI, stripped of the hype, is a tool for which demand will continue to grow strongly. As demand grows, we seek the type of exposure that is diffused through a range of different companies along the AI value chain.
Investing in the global equity disruption thematic requires looking past the next few quarters when setting portfolio weights. This is particularly so with AI where it may take many years to roll through companies in a range of industries.
We believe that the AI thematic still has a long way to run. However, investments must abide by strict valuations and where they fail to do so, capital will be moved towards those opportunities that are not yet fully priced.
Share this Post