Monday
The week starts rather more conventionally than it is going to end. There is the usual portfolio housekeeping: monitoring the funds, looking at performance and positioning, dealing with client queries and catching up on what has changed across the companies we own.
I also spend part of the day in internal stock discussion meetings. One of the benefits of working in a global firm is that you rarely look at a company in isolation. A discussion about a semiconductor company in Taiwan can quickly involve colleagues covering technology elsewhere in Asia and the US, particularly when the underlying debate is something as global and relevant for portfolios as AI infrastructure.
The rest of the day is spent preparing for the week ahead. I am travelling to Taiwan to meet companies across the AI supply chain on their own turf, to better understand the key moving parts of what is a very dynamic sector, before heading on to Hong Kong for a broader Asia company conference.
Preparation is probably the less visible part of company meetings, but often the most important. I want to arrive knowing where our assumptions differ from the market, but also with an open mind about where we might simply be wrong.
Tuesday
Most of today is spent travelling to Taipei. I use the time to finish preparation and go back through research on the companies we are meeting.
A trip like this is particularly useful for an area like AI because no company operates independently. A semiconductor designed in the US may be manufactured, for example, by TSMC, and packaged alongside high-bandwidth memory, incorporated into a server and connected through networking equipment, with dozens of specialist Taiwanese companies involved along the way.
The aim is not just to learn about each company individually. By asking similar questions at different points in the supply chain, you can start to work out where the answers reinforce one another, where they contradict, and occasionally where something you thought you understood isn’t quite right. There can also be useful read-across for the US technology companies that now make up such a large part of global equity markets.
Wednesday
The first full day begins early and includes meetings with Jentech, Accton and Nanya, followed later by SINBON and Fositek.
It is a useful cross-section of the technology supply chain, from components used in increasingly complex semiconductor packages through to networking equipment, memory and cooling infrastructure.
One theme quickly emerges. For many of these companies, the AI opportunity isn’t simply about selling more. As AI systems become more powerful, they require better cooling, more complicated components and increasingly sophisticated semiconductor packaging. That means some suppliers can potentially provide more valuable content in each system rather than relying purely on higher volumes.
Our subsequent work highlighted companies that could benefit from this increasing content and complexity while also improving their competitive position.
That distinction matters to us. It is easy to get excited about companies growing quickly when demand is strong, but we are ultimately looking for businesses where that growth can translate into sustainably better economics, rather than simply benefitting from a short-term bottleneck in demand and supply that will eventually disappear.
Thursday
Another early start as we travel to the south of the island for a particularly important day of meetings, culminating in a meeting with a company that manufactures many of the advanced chips designed by companies such as Nvidia. It is the standout meeting of the trip.
I went into the meeting wondering whether its competitive position could really get much stronger from here, particularly with competitors spending heavily to catch up. I came away feeling that it might.
What was particularly striking was that almost every part of the ecosystem seemed to lead back to this particular company. Customers, suppliers and packaging companies build their own plans around its technology roadmap. The complexity and ecosystem it has built in Taiwan around leading-edge chip manufacturing represents a competitive advantage that is difficult to replicate and isn’t something that can be recreated simply by building a new fab elsewhere.
Our view on the company was sufficiently strengthened by the trip that we have subsequently added capital to the position.
One other theme keeps cropping up: the sheer amount of physical capacity that needs to be built to support the AI boom. Across our work on our future supply chains strategy, we have seen manufacturing footprints becoming more diversified across Asia, and the meetings this week reinforce that trend.
For all the focus on chips and software, AI ultimately relies on factories, equipment and infrastructure being physically built. One of the nearer-term risks may simply be whether the industry can build everything being asked of it quickly enough.
Friday
Our final day in Taiwan includes Elite Material, ASE, Compeq and Gold Circuit, companies involved in the substrates and circuit boards on which AI chips from companies such as Nvidia ultimately sit.
By now, patterns across the previous three days are becoming clearer. One is the difference between structural winners from AI and companies enjoying a cyclical upswing. Some businesses are benefitting because their products are becoming more complex, valuable and strategically important. Others are benefitting from shortages or unusually favourable pricing.
Separating the two is crucial. Strong earnings growth does not necessarily mean the underlying business has become better.
Travelling between meetings also gives me time to keep on top of the day job. Markets do not stop because you are on a research trip, so there are portfolio questions and client requests to deal with between meetings.
On the late flight to Hong Kong, I spot the large offshore wind turbines in the Taiwan Strait. During the week, locals had mentioned an interesting potential strategic dimension to Taiwan’s extensive offshore wind development: as well as generating renewable electricity, all that physical infrastructure in the Strait could potentially also be of broader resilience planning.
Saturday
Saturday provides some welcome breathing space after three packed days of meetings. I take the opportunity to hike up Victoria Peak with an old friend who now lives in Hong Kong, providing some much-needed exercise after a week largely spent sitting in meetings, planes and restaurants.
I go back through my notes and compare what companies told us. This is often where the real value of a research trip emerges. An isolated management comment is interesting, but hearing the same message independently from companies at several different points in a supply chain is much more powerful.
As I write up my notes, I start thinking about what, if anything, the trip should change in our portfolios. A couple of existing holdings look more interesting than they did at the start of the week, and one potential new idea has made its way onto my research list.
Sunday
Sunday starts with the rare luxury of a lie-in, with no one or three-year-old waking me at the crack of dawn, much to my wife’s envy back home.
By the evening, it is back to work with a conference dinner, although catching up with current and former colleagues over dinner is a really enjoyable part of the job. It is also a chance to compare notes before another packed week of company meetings, this time extending well beyond the AI supply chain.
The week sums up much of what I enjoy about fund management. One day you can be buried in financial statements, quarterly results and internal stock debates; a few days later you’re sitting across the table from management on the other side of the world trying to work out whether what you believed at your desk still holds up in the real world.
You rarely come away with all the answers, but hopefully you come away asking better questions. Ultimately, the job is to keep challenging what you think you know, work out what really matters and decide whether the evidence is strong enough to act.
