The Delphi 2025 Annual Update reviews portfolio positioning, market conditions, and the key investment themes that shaped the year, including tariffs, economic efficiency, and artificial intelligence.
The year began with declines following tariff policy, but stock prices recovered quickly.
Toward the end of 2024, we adapted the portfolio accordingly, selling shares of companies we believed were exposed to tariffs and focusing on companies whose operations allow for progress and efficiency even during periods of economic slowdown, should such a slowdown occur as a result of tariff policy.
These are companies that lead in growing fields, contributing to improved economic efficiency, whether through the accessibility of computing resources, more efficient commerce, more effective advertising, or more efficient logistics infrastructure. These companies contribute to growth by making it easier to establish and operate businesses, and in times of slowdown, companies will look for ways to become more efficient, and these services enable them to do so.
In addition, we are careful to invest in financially established companies with exceptional resources. If we were indeed to encounter a period of slowdown due to tariff policy, these companies would be able to use their large cash reserves to acquire less resilient companies and strengthen their competitive position.
In practice, the economy did not experience a recession or a significant slowdown. Tariff policy softened and stock prices recovered, with a large part of the driving force of the economy being the continued demand for more and more computing resources. ASML and APH, two companies that were added to our investment portfolio at the beginning of the year, benefited from this trend and contributed to portfolio returns, as we showed in previous letters.
In principle, and especially toward the end of the year, we prefer to look forward rather than backward. The rise in share prices in our portfolio was accompanied by growth in the profits of the companies in which we invest. We estimate that the pace of profit growth should remain similar in the foreseeable future, and therefore we do not expect changes in the return potential that our portfolio can generate.
Artificial Bubble?
Artificial intelligence applications are becoming increasingly common and appear to be here to stay. Alongside this, there are growing concerns about the ability of companies such as OpenAI (the owner of ChatGPT) and others to make money. Indeed, although these are private companies, the data show that they are probably losing enormous sums that continue to grow as the companies expand. Therefore, it is very legitimate to ask: is it even economical to develop such models? Are we witnessing the formation of a bubble?
On the one hand, yes. The ease with which companies in the artificial intelligence field raise money is reminiscent of the dot-com era. “Circular” transactions between NVIDIA and its customers, which we wrote about in letters to investors during the year, also raise doubts about these companies’ future ability to continue purchasing chips and computing services.
On the other hand, large corporations such as Google, Amazon, Microsoft, and Meta are also active in this field. Beyond the fact that they earn enough money to finance the development and operation of artificial intelligence models, they do so at a lower cost. Google is probably the best example: it develops and runs artificial intelligence models on chips that it designs itself and does not need to pay the high prices charged by NVIDIA. Beyond the fact that Google’s chips are cheaper, they are also more energy-efficient, further reducing costs. In addition, Google is a provider of computing infrastructure and does not need to purchase such infrastructure from third parties.
The fact that these four giant companies can develop chips for themselves hurts NVIDIA’s future growth potential on the one hand, but the result is that they can provide artificial intelligence services at a significantly lower financial investment compared to smaller companies.
OpenAI and other companies may possibly go bankrupt at some stage, but even if that happens, the four giants we mentioned will likely continue to provide computing infrastructure and artificial intelligence tools to their customers.
Delphi 2025 Annual Update – Summary and Outlook
As detailed in the Delphi 2025 Annual Update, our strategy continues to focus on high-quality companies with durable growth and strong financial foundations.
We continue to focus our investment portfolio on the highest-quality companies we can find. They are leaders in their fields, consistently increase their sales and profits, and have enough cash to weather periods of slowdown and recession. Stock prices in the market may rise or fall over short periods of time, but over the long term, these are the parameters that will allow us to continue generating returns consistently.