It’s a pleasure to provide you the November portfolio update! Let’s dive straight in.
Performance
The portfolio is up 5% year to date against 13% for the S&P 500 in EUR. After 18 positive quarters, Q3 was the first negative one, and Q4 is down 4.4% so far, in line with the S&P 500.
The portfolio is currently not available live to subscribers. I’m afraid it would encourage copy-paste trading and expose you too directly to my mistakes.
Subscribers have done well doing their own analysis of my write-ups though, which are up 19% on average since pitch, against 9% for the S&P 500 over the same periods.
Recently many of the older pitches have lost some feathers, as gold has topped out and some of the construction names are hit by higher interest rates and oil prices.
The portfolio updates helped handle this in real time, although I am still holding Lincotrade and Generalfinance, which also came down significantly, for company-specific reasons.
October 2026 market events
The S&P 500 made a new all-time high on October 6th, but ended the month 2% lower. The Nasdaq lost 6%, the Russell 2000 only 1%.
Anthropic filed their S-1 to go public. Surprisingly, this showed that they only put about 6.7 billion dollars of net cash into the company since 2023 to build arguably the most impressive tech business around today. This includes capex of 1.1 billion.
Almost all compute is rented and cash operating losses have been very limited. Below is the cumulative sum of investments and operating cash outflows since 2023:
When I started digging a bit deeper, I became even more intrigued. The S-1 splits the costs by function, note 8 on page F-40:
Over 60% of operating expenses are not pointed towards “running the business”, but rather at developing and training the best model.
Below, I have put this into their income statement, variable and fixed costs separately, and estimated 2027 and 2028 revenues, with growth coming down from 1000% to about 100% on the current run rate in 2027 and 50% in 2028, and just theoretically with fixed costs (model development, overhead, …) constant at today’s level:
The operating leverage is huge. Development, marketing and overhead are fixed at about 52 billion a year. If Anthropic kept them there, continuing to improve the models with today’s resources, and revenue grew to the level Anthropic itself projects for 2028 according to Reuters, net income would be about 50 billion in 2027 and 94 billion in 2028.
But like Amazon in the 2000s, Anthropic invests through the income statement, expensing the model training costs. Under IFRS most of the development would be capitalised: Anthropic would carry about 14 billion of models as intangible assets, amortised over four years, and would have made about 7 billion of operating profit over the last twelve months, 5.5 billion after tax, on an average invested capital of about 12 billion, a return on invested capital of roughly 45%. The 2027 and 2028 net income above would be 61 and 101 billion.
So this IPO might be relatively cheap: at the 1.5 trillion talked about, it would trade at about 30x 2027 and 16x 2028 earnings in this scenario. But wait a second, do these guys actually have a moat? That question will ultimately decide whether the AI companies keep any of the benefits of their technology, or whether 100% of them are reaped by the users.
Every single player in the AI landscape might look a lot like a telco player soon. They provide a very useful service, have high capex, are undifferentiated, and ultimately achieve a terrible return on capital, with massive benefits going to users of the technology.
I think AI will create a tsunami of corporate profits, but it’s unclear if most of them will land in Nasdaq or Dow Jones type companies. This makes the Anthropic IPO, and the AI company landscape interesting to debate about, but less interesting to invest in, from my point of view.
It’s more interesting to me to pick out unique and undervalued micro-caps, at least while AI is not scooping them up before we do… On to the portfolio now, without further ado…
Changes and Portfolio
In this section:
Full portfolio with % allocation and change vs last month
All portfolio moves and their rationale