Monthly Archives: February 2026

Some links 06/2026

AI Entrepreneur Matt Schumer on how AI changed the way he works as a Software developer (TwiX Article)

Some deep thoughts on how AI could change the job markets using examples from history

A great comment on Volkswagen’s recent “Cash flow surprise”

The Brooklyn Investor is back with a sanity check (Market PEs, AI, Berkshire)

A great article on a version of the “PE Secondary Shenanigan” at Ares

French Wine and Spirits exports have been suffering in 2025 and hit a 25 year low

So far, Radiology has proven to be more or less immune against disruption from AI

Quick Updates: Frosta, Alimanetation Couche-Tard, Bombardier, Bouvet & Robertet

This week was quite busy, with (at least) 5 of my companies releasing 2025 numbers, some more preliminary than others. Overall it was a “mixed bad”. Some good, some not so good. Let’s jump in:

Frosta

Frosta made a premliminary earnings announcement that was clearly below my expectations and at the lower end of the guidance. Thanks god, I covered my Axx with that statement in my write-up from earier this week:

“One important thing to understand with Frosta is that they don’t try to smooth earnings much, at least not to the upside. So you might always be in for some surprise either at the 6 month mark or annual report. Their guidance is usually very wide. Sometimes they decide to increase marketing expenses significantly which lowers profit in the current year but boosts revenues in the next year. In the long run, this has turned out very well for shareholders but for “weak hands” this can be a little bit unnerving.”

I guess this was once again such an event. Net income actually declined -12% in 2025. On the other side, growth of the Frosta Brand with ~16% and the ready made meals with almost 18% in 2026 is far above the market growth. So Frosta is taking significant market share. The dividend will remain at 2,40 EUR. They guide towards overall growth between 4-9% and a net margin of 4-8% in 2026 which is of course again very wide but implies that growth will continue for the Brand.

According to the just released annual report , my Quick and Dirty analysis looks as follows:

Frosta AG – Defining their own Category the Hard Way

Disclaimer: This is not Investment advice. The stock discussed is very illiquid and trades at the unregulated market (Freiverkehr). If you want to buy this stock, work very carefully with strict limits. The author owns the stock and might buy/sell it without giving prior notice. And as always: PLEASE DO YOUR OWN RESEARCH !!

After having two (relatively) exciting stocks in the last two weeks with Rocket Internet and Innoscripta, I decided to tune down the exitement a little bit and focus on a very boring, family run German small cap this time. In order to not fall asleep, you might want to listen to the Soundtrack while reading:

Elevator Pitch:

This write-up is special in two ways. For one, I have privately bought the stock already a few months ago. Secondly, I base this write-up on another write-up from my friend Jon from abilitato.de. So please read that one before you read my “mini-writeup” where I only focus on a few specific aspects.

In a nutshell, Frosta is a boring, under-the-rader German family owned and run frozen Food company that does not do a lot of investor relations but runs a very convincing strategy focusing on additive-free ready made frozen meals. Inventing this category more than 20 years ago, the main Frosta brand is now growing with solid “mid teens” percentage rates p.a., has succesfully managed to enter and grow in neighbouring countries and with profitability that is steadily increasing. For the quality of the company and the potential growth prospects, the stock is still relatively cheap in my opinion at around 12×2025 P/E (ex net cash).

Here ist the write-up. Best read it after having a decent “Chicken Paella” from Frosta 😉

Some links 05/2026

Ben Thompson (Stratecherry) on Microsoft and Software in general vs AI

Larry Swedroe explains nicely why IRRs for PE funds can be so misleading

Some thoughts on AI, Software and the impact on Rightmove

Contrary Research’s annual Tech report is worth reading

2025 letter from LMN Capital (in German. Biontech,Intred, Femsa)

Very deep thoughts on the AI risks of a data driven company like RELX from the Fiercely Independent Investor Substack

A nice write-up on the history of cross border payments and the business model of Wise

Quick Update Innoscripta (Company Conference Call)

My timing of my Innoscripta write-up was a little bit unfortunate. Just a day later, Innoscripta held a conference call explaining the 2025 numbers. I think I need to better check the calenders of the companies I write about in the future….

The presentation can be found here.

I listened to the Call on Quartr. My main (and of course subjective) take-aways:

  • as a relatively new initiative, they develop a “safe storage” for R&D data
  • the arguments about LLMs and their limitations did not overly convince me. The arguments rely on today’s abilities of LLMs, but the question is how this will develop in 1,3 or 5 years. Especially I do not see that the LLM in the hand of a client itself is the competitor but competitors (especially other consultants) using AI to rapidly develop (cheaper) alternatives
  • Cash conversion (to EBIT) is currently around ~60% but they try to improve cash collection
  • They are bullish on the Government increasing the programs in general (regulatory tailwinds)
  • They claim that they see increasing revenue from the same clients which I find quite surprising. They didn’t however provide any concrete numbers
  • The clarified that the overall proces from application to cash collection is 9-12 months (3 months for innovation approval, 6-9 months for tax credits)
  • They claim that the “front loading” of 4 year applications seems not so significant
  • There is a clear seasonality that companies hand in filings mostly in Q4 in order not to lose credits for T-4 projects
  • Revenue in Q1 2026 should look good (whatever that means, not sure if they compare this to Q4 2025 or Q1 2025)
  • The currently do not consider M&A for international expansion but try to grow organically in 4 different countries (UK & US was mentioned)
  • The question if the 40% growth for 2026 is relevant was somehow confirmed but very indirectly. A “real” guidance should be expected for the AGM in April
  • Interestingly, the overall growth in the German Tax Credit market was +60% in 2025 which means Innoscripta’s growth is in line with the market and they want to grow with the market. More details again shall be provided at the AGM. The analyst mentioned that the market will grow less than 40% in 2026
  • The question how big the “4 year front loading effect” actually is, they were quite evasive. They mentioned that there will be growth internationally and through “Software” and Tax Credits is described as a “Milestone” or “trojan horse” to get other business.
  • Cash usage: No additional cash is needed for growth. Buy backs limited due to limited liquidity but rather distributions.

Overall my impression is that at least 2026 should look pretty Ok, but further out it gets a little bit vague. International expansion is clearly more risky than increasing market share in Germany.

But overall it remains a very interesting and dynmaic company and I am looking forward to the Annual Shareholder Meeting which I hopefully can attend.

Innoscripta SE: 60% EBIT margins & rapid growth but why did the 2025 IPO flop ?

Management summary: 

Innoscripta, a young German “SaaS company” which IPOed in 2025 came to my attention because it is extremely profitable (EBIT margin 60%) and growing like crazy (10x in sales since 2020). However, because of the unique revenue model (success fee instead of software fees) and a rapid decrease in quarterly growth in 2025, I am currently not investing, although the stock is not super expensive at 21x trailing P/E. But I will keep a close watch. 

The 2025 IPO 

Innoscripta was one of the few “official” IPOs in Germany in 2025. If we look at the share price, it was not a very successful one: 

Read more

Some links 04/2026

Szew’s Substack with an interesting write-up on Wehosters Ionos & GoDaddy

Once insider buys are published, there is little Alpha left according to this study

The Slow Compounding Substack with a very interesting comparison between B&M retail and Dino Polska

Interesting long read on the history of Waste Management, the US company.

Annual letters:

Whatever you wanted to know about the “State of the art” of batteries is in the 2025 report from the Volta Foundation
Google DeepMind has released Project Genie 3, an AI tool that makes it easy to create playable 3D worlds by simple prompts. Video Game stocks took a direct hit.