This is actually great humour

Michael Mauboussin’s latest paper calculates a CAP value for Microsoft of ~18 years.

In other words, Microsoft will deliver excess value from its strong moat the next 18 years.

Given its size and the European sovereignty theme this scenario seems less and less likely…

Ah…I was not precise here. With Gemma 4 it looks to me that we are headed in that direction. Maybe the 99/1 is a stretch but I think local models will dominate the bulk of queries over time on phones.

Not today, it is a prediction for the future. Personally I never do Google search any longer.

Google has low transparency in reported figures so it is difficult to assess how the old search business is doing.

It is widely reported that AI Overviews has reduced traffic from search to websites.

Jeg har en anden hypotese. Unge sparer markant op fordi huspriserne galoperer afsted og dermed skal der akkumuleres kapital hvis man skal ind på boligmarkedet.

It is 100% where we are going and the view that Apple has taken. You can probably solve 99% of all queries with a local LLM on chip and for the remaining 1% you route to an AI datacenter.

Google’s search business dies but instead collect royalties for installed Gemma models on iPhones.

Makes sense Mikael.

I’m wondering whether there is also an anchoring effect so 7% mortgage rate today feels terrible because of the previous long period with very low mortgage rates vs 8% in the 90s rising from a normal average.

US economy is growing close to trend so things are not that bad.

The current values do not reflect actual behaviour in transaction data, so there is a disconnect.

From a macro perspective I put very low weight on these surveys.

We have seen similar pattern in Denmark for a long time. There is something structural happening in these surveys.

My guess is that they over time have been more correlated with political feelings and narratives, and social media feeds are generally biased negatively which have impacted surveys.

And then you have Oracle down. The market is quite inconsistent.

The meltdown in software and cybersecurity is getting ridiculous.

Fund managers are completely panicking.

Google gets a new money machine as its old search business dies slowly over the next decade. Beautiful.

Was Apple smarter than everyone? Instead of burning hundreds of billions of dollars it will partner with Google to get Gemma 4 and future AI models to run locally on the phone.

Apple gets good enough AI on their phones and can avoid massive CapEx. It will end up paying royalties to Google.

Software went from AAA business to something AI optimists believe will almost go out of business over 1 year.

I have never seen anything like it before and surely we are in the phase where EPS junkies and index hugging fund managers are giving up…

I also don't care about dividends. Buybacks are more efficient, but those dividends have a fan base ;-) And Adobe could use a new group of bullish investors. I have also looked at the outstanding shares and thought a couple of times "wait a minute"

When even Palantir is sunk by Anthropic nothing will stand.

What do you do if you are Adobe?

You lost the growth narrative in the AI era and thus your previous shareholder base left for other companies.

Do you only aggressively buy back shares (which they are) or do you also add dividends into the return of capital mix to attract new investors? $ADBE

This exactly why Google/DeepMind is likely the long-term winner. Demis Hassabis has deep knowledge of the topic developed over a very long period of time. $GOOGL

Wrong.

Not because of self-servicing model but because of massive change in the franchise model selling company-operated stores to franchise owners removing employees from the direct operating model. $MCD

Given that the 10 largest stocks account for ~40% weight of S&P 500 and their economic value added profits are ~66% of the index, it tells you that the market expects their ROIC to decline over time as a function of their massive CapEx and likely lower ROIC on AI infrastructure.

The potential issue with this analysis is that it likely does not account for confounders - one could be that technology companies overhired and thus were going to reduce employment regardless of ChatGPT launch. It is a classical mistake when doing causal analysis.

Han har jo længe talt om “kakerlakker” i køkkenet.

JPM har modsatrettet interesser her. På den ene side så giver de store kreditter til private credit (PC) firmaer og har dermed upside i PC.

På anden side er PC en form for skyggebank-aktivitet (konkurrent) og dermed modsat reguleret banker som JPM

They are not without humour in Spain

24.7x 26e EBIT is totally lack of margin of safety given the state of the US consumer.

And yet it remains quite expensive in a very mature market.

3) European technology sovereignty will expand faster. Microsoft, Google and AWS are most at risk from this emerging trend.

4) US equity valuations will over time reflect a growing risk premium vs rest of the world. Chinese/EM equities will slowly win larger share of global equity flows.

#equities

1) The electrification adoption curve will accelerate (EV, heatpumps, and power). Solar and battery will grow even faster challenging the nuclear renaissance before it even started.

2) European defence investments will accelerate at an even faster pace with EU defence stocks winning big time.

The war in Iran is still largely unpredictable and as a consequence all first-order effects. Everything can change on a dime it seems.

The more stable outcome of this war and its longer term second-order effects, which are far more interesting for investors are in my view as follows:

Enig, prisen med eller uden afgifter vil jo ramme det niveau hvor markedet går i ligevægt.

Just like the Sora model, we will come to the conclusion one year from now that cybersecurity is vastly more complex than scanning scripts for vulnerabilities.