DISCLAIMER: NOT INVESTMENT ADVICE / DO YOUR OWN DUE DILIGENCE

Company performance

My portfolio’s annual earning power per my estimate represents around 5% of my portfolio value, i.e. they are attractively priced at 20x their intrinsic earning power. To understand this “intrinsic” concept, in a world where these companies do not reinvest earnings into creating bigger and better things I would expect the annual dividend I would receive from my companies to yield roughly 5%.  

But in reality the companies do reinvest their earnings.

At the core of it, I would prefer the companies to do such a thing as long as such reinvestment generates incremental earning power in the future.

More often these incremental earning power do not show up on paper in the short run (e.g. Meta, Amazon, and Google will spend c.$100bn worth of annual capital expenditure for AI- and cloud-related projects that may not yet generate returns in the immediate next year). As long-term investors I however do care about how they are going to show up in the long run. 

At 20x their intrinsic earning power, they look reasonably attractive to me, as I believe these companies are better than an average S&P 500 company, and such an index is currently valued at 29x (source: https://www.multpl.com/s-p-500-pe-ratio).  I also believe that they can grow faster than the S&P 500 average company. 

Artificial Intelligence – Large language models

This quarter, I have primarily focused on experimenting more and more with LLM.

When ChatGPT was released on November 22, I started testing it a week after its launch to assess its usefulness. I had my own “ChatGPT moment”. I was amazed by the user-friendly nature of chatbots as the first practical interface for large language models (LLMs).

Today, I use LLMs to help me write codes from scratch. As someone who was not previously proficient in any shape or form in coding, ChatGPT has enabled me to generate functional programs on my own. This capability makes me believe that, in the future, advanced versions like ChatGPT-6 or ChatGPT-7 could allow anyone to create their own software simply by using natural language prompts.

This concept aligns with the idea of AI factories, where AI autonomously writes programs and software to generate intelligence, as noted by Jensen Huang.

Who’s in it?

The companies leading in the AI space are Alphabet, Amazon, Meta and Microsoft (the former 3 are in my portfolio). Each of these tech giants operates within different layers of the AI stack, leveraging their unique strengths.

Source: Stratechery

Google stands out as the only company independently pursuing AI across all layers. This is an ambitious endeavor that may stretch the company’s resources. However, as the pioneer of the Transformer model, which underpins LLMs, Google may just know what they are doing. In a field that changes almost every day/ week / month, just roughly knowing what one is doing probably is a great competitive advantage already.

The implications of these varied AI approaches remain uncertain. The question of where the initial value of AI will accumulate is still open, with many suggesting the infrastructure stack at the moment. This perspective is supported by Nvidia’s stock reaching new highs. But we’re watching patiently to see where the ultimate value would accrue to.

Other notable updates

Prada released earnings during the first quarter of 2024, showing Miu Miu significantly outperforming others with 90% year-over-year growth, topping the brand heat index measured by Lyst. Prada maintained low double-digit growth. This growth should theoretically improve EUR/square meter metrics since the company hasn’t expanded its square meter footprint. Estimates of EUR 20k/sqm going into the quarter suggest Prada may achieve EUR 23k/sqm this quarter. Post variable costs and direct sales costs, this could result in a +300bps margin expansion unless management chooses to reinvest elsewhere.

Unilever has seen long-awaited volume growth over the past two quarters. The current management is revitalizing the company by aligning management and workforce incentives with performance goals, market share gains, and incremental profit targets.

I will discuss the rest of the other positions (brookfield, blackstone, new york times) in the coming quarters.