Read more

Read more

Let’s start with the elephant in the room: Closure of the strait of Hormuz. I think this is a non-event and a no-brainer opportunity to load up if you are focussed on the medium to long term (say 6-12 months+ out). Only issue might be timing, so I did move to some cash (sold Inmode, Directa Sim, Humble group and ICCC) as they were my weakest positions.

I have seen some panicking about this and the best argument against panic is that you would need Iran, a $400 billion economy who is massively underperforming expectations in the war so far, to basically nuke the global economy. And that somehow militaries with a budget larger than Iran’s GDP will find no way to actually get those tankers through.

Some observations:

• Trump isn’t gonna TACO his way out of this. He can pull out, but that will only worsen the situation. Iran has to TACO basically.

• I think oil might briefly spike to $200+ before this is over. As clearing the strait will probably not happen before early April.

• The market kind of trades like an option here. Stocks are slowly going down, oil is slowly crawling upwards everyday this is not resolved. The moment it is resolved (either a deal or strait is effectively cleared) oil will crater down, and the market will spike upwards.

• I have seen posts about how ⅓ of global urea supply has to go through the strait. Except it is more like 9% of global production. About 18.5m tons with 200-240m tons of global supply.

• Oil stocks are just as/more expensive than in the 2022 peak. Petrobras is trading far above its 2022 peak, most of US shale producers are trading near 2022 peak levels, you would really need several years of $90-100 oil to see modest upside here. I remember buying Petrobras preferred shares in late 2022 for $9 and I got $3 of dividends within a year, now they are at $17 with a lower dividend payout rate.

• A similar % of oil went through the strait during the Iran-Iraq tanker wars in 84-88 in which over 200 tankers were struck. Iran also attacked Saudi and UAE ships back then. It barely seemed to have affected oil prices (not inflation adjusted) and US only intervened in middle of 88:

• The US has become A LOT better at these bombing campaigns. They had real problems finding even a minority of Iraq’s SCUD launchers in the first Gulf war in 91 even with spec ops teams on the ground. But have no problem now taking out Iran’s BM launchers with over 70% taken out so far.

• If the IRGC is rational they put a reformist leader in place and pledge no nukes, no ballistic missiles. Trump would accept that and force Israel to back down. If they do this before the US manages to open the strait of Hormuz, they have some leverage (threat of a future closure) and can stay in power. If the US manages to open the strait the IRGC loses all leverage and this likely becomes background noise and Israel/US will try to continue until IRGC + Basij is completely hollowed out.

• Unlike Libya or Afghanistan or even Iraq there are actually functioning state institutions + army and a fairly unified country humming below the IRGC + Basij oppression system. So I actually have good hopes that Iran can recover nicely here if those 2 organizations are hollowed out.

• When this is over, Gulf states will build pipelines to bypass the strait and get better missile defenses. So again, IRGC has some room to make a deal here, but in 2-3 years will have lost most/all their leverage.

• Global supply of crude that has to be rerouted is also a lot less than 20mb/day. Actual crude is like 15mb/day, of which:

  • 1.5mb/day of Iranian crude

  • 5mb/day rerouted through various pipelines

  • Probably 1mb/day of supply elsewhere in the world can be added within 3-6 months

  • Another 1mb/day of demand destruction at $100+.

So the real amount that has to be squeezed through is about 6-7mb/day.

• 1 VLCC is 2mb/day (probably a bit more complicated though as there are a wide variety of oil + product types that cannot fit all into a few tankers).

• That said there is some tail to IRGC’s missile launches:

  

• Shooting down Shahed drones in a cost effective way is already a solved problem.

• The IEA countries have 1200mb, China has 1300mb of crude reserves. Probably another 1-2000mb/day in the rest of the world in crude reserves. Throughput might not be high enough to backfill that 6-7mb/day though.

• Asia gets ~80% of oil that flows through Hormuz. You would think that holding a gun to China’s (Asia in general) head on this would be a bad idea for Iran strategically.

• Greek shipowners have also snuck 10 oil tankers through so far.

• In 2019 dollars, oil is currently sitting at only $75/barrel. January 2027 futures sit at about $55 in 2019 dollars.

• The oil futures market is possibly manipulated though. So if you want to bet on this, just buying Brent futures would be the best way. The problem is that it is fairly likely something will be negotiated/done in the coming weeks and your futures gap down to $70.

• Iran closing the strait has been war gamed for decades. So take all those claims that this caught the administration off guard with a very large pinch of salt.

• In the medium term, Argentina, Venezuela and possibly a free Iran will increase production by millions of barrels/day while 1000km EVs cheaper than ICE cars are in the pipeline. So I don’t see how anyone can be bullish on oil beyond the next 12 months.

That is my 2 cents, I don’t see any attractive ways to bet on this other than just buying stocks that are now getting cheaper because of this. My strategy has been to sell down stocks that I found less attractive and slowly get fully invested in the coming weeks. The best time to buy oil stocks in 2020 was when oil went negative. The best time to buy euro stocks was in 2022 when inflation was peaking.

Read more

Read more

Read more

Read more

Read more

General observations:

• LLM models are rapidly becoming commodities for most use cases (I have said this before several times, but bears repeating).

• Some observations from extensively using various paid models for various types of tasks:

  • I think longer deeper thinking modes beat raw model capability. Currently I am only using OpenAI’s GPT 5.2 with extended thinking for research/learning new things. Yes it sometimes takes 5-10 minutes, but you get far superior answers vs quicker Gemini and Claude thinking modes. I wonder what relative performance would be vs for example Gemini if it thought longer?

  • It does seem that when you go back and forth between skeptic/confirmation bias mode, LLMs converge towards the correct answer/state. This implies that using more tokens for a lot of tasks could still greatly enhance performance without improving models themselves.

  • I think a lot of people make the mistake of using some quick version of a model and then when the answer is crap they think AI is crap.

  • The above implies that much cheaper tokens that provide answers at greater speed could disrupt a lot of industries/jobs in the coming years.

  • LLMs for researching stocks work best when you do most of the thinking and LLM does most of the fact collecting/summarizing from a variety of dense sources.

• Proprietary data advantages mostly work when it is messy text/visual/audio data that needs to be quantified/mined for insights. If it is already nicely labelled quantitative data I don’t think recent AI breakthroughs really help much vs preexisting machine learning techniques.

  • That said, I think huge fortunes could potentially be made by sleepy companies who do possess this data, and that is where the big winners are. But IMO try to argue explicitly and in great detail why these LLM models will help them and what the advantage is.

• According to Huang’s law GPUs double in performance every 2 years. That means in theory in 10 years you get 32x the performance for not that much more cost.

• This has some implications:

  • I think some tasks are currently not practical for LLMs for most people because slow speed and cost, a 32x cost decrease will change that.

  • Somewhat mitigated by fact that tokens are currently heavily subsidized (there will be price increase, but on cheaper token cost)

  • This does not mean power consumption goes up 32x. It may go up 5x? 10x? I don’t know. A large part of the performance increase will come from things other than smaller transistors. For gaming GPUs watt/frame drops quite a bit for newer generation GPU due to smaller transistor size. By almost half over 2 generations.

  • Bearish on Data centers though. Local energy needs will go up, so that could become a bottleneck as physical space needed for 1 unit of computation declines by a factor 32x.

• Those scores/tests these LLM companies like to brag about often mean jack shit.

• So what advantage do the major LLM companies have:

  • Economies of scale when it comes to buying/renting compute

  • Being a user funnel towards various services (kind of like Google)

  • Proprietary data from questions and user feedback from chats

• It is good to be reminded once again that specialists usually beat generalists in the game of capitalism. LLM companies are very much generalists. So take “Gemini/ChatGPT will eat x’s lunch!” narratives with a grain of salt.

  • With some exceptions for example in investing where sectors can become overpriced/underperform for longer periods (energy specialists will almost always underperform vs good generalists who know their limitations).

• Especially when LLM giants do not possess a lot of proprietary niche data themselves.

• Companies with right proprietary data will likely have their own LLMs that will end up beating Openai, Claude etc within their niches. They use their data to create a superior experience within the niche, gain/keep more users, and get in turn more proprietary data from user chat history.

• For niches LLMs do not have much of a size, funnel or user data advantage unless dominant players in those niches stay asleep at the wheel.

• Content generation players will either be wiped out or become an even stronger, more prolific signal beacon in an ocean of slop. For example I can see already at least somewhat trusted news providers becoming a more dominant safe haven with prolific deep fakes etc.

• This one is true for software: building it was often not a bottleneck before LLMs BUT:

  • I do think that demand for software still vastly outstripped ability to supply it in especially niches. Lots of things were not viable to build simply because hiring an engineer for $80k wasn’t worth it given the small user base (and learning how to build had too steep barriers).

  • Some somewhat inexperienced engineer/employee may have previously thought “wouldn’t it be cool if we have this feature?”. But may have been stopped by the fact that getting a bunch of different experienced enough engineers together to build it was too much of a barrier.

  • Established players however have a data advantage here. Current providers (like Constellation) will already have their client relationships and know client work flows better than most. If they are proactive in finding these small niches that are now cheaper to provide they would still have the advantage here in simply upselling new features.

  • The risk here is that engineers/employees at those companies will also know this, and if not proactive enough will leave and found their own companies and eat away at the edges here. The barrier is much lower to do this now.

  • Switching costs will come down.

  • But that could also mean dominant players with superior products/services will eat more market share from inferior players that were protected by these switching costs.

  • Tech debt will become a huge problem for companies that use these LLMs carelessly and buy into the agent swarm hype.

• When it comes to writing software:

  • Either you write it yourself/do pair programming understanding most of it (the specialized/complex/niche part). Not that much productivity gains here for experienced engineers, but will lower barrier for less experienced engineers to build a wider variety of things at greater speed.

  • Or you do it with agents understanding close to 0% and using tests to verify it works.

  • Not much in between as fixing/trying to understand messy agentic created code is a huge time sink.
    From experience and talking to people who work in software it still makes huge structural mistakes. It is terrible (and has not improved that much with newer models) at big picture stuff like putting code in the right places, writing DRY (don’t repeat yourself) code, allowing impossible states, creating useless unit tests.

  • The unevenness in quality is a problem. A somewhat experienced engineer might not produce the greatest code and actually be outperformed by Claude in a lot of cases, but they will also not randomly every other week create truly terrible code (like these agent workflows tend to do) that will add huge tech debt with little code.

• I think LLMs remove a lot of friction especially for people who are not that computer savvy (which means a lot of boomers who decide on large budgets). Previously if you wanted to know something you had to type in the right words in google and then read a bunch of websites. Or ask on forums. So businesses that benefited from this friction will be in real danger.

Will LLMs break network effects? By aggregating data? Honestly I think each network effect/industry has to be assessed on its own as they are all different. But let’s take Rightmove as an example as they sold off hard. Real estate agents pay them a double digit % of profit, so there is a pretty strong incentive to find alternatives (they control 80% of the market and are a de facto monopoly).

I started looking for houses recently and on the consumer side I can very easily see a much more improved experience. For example:

• Show me a map with $/sqft and/or other constraints other than price. Rightmove doesn’t even have this.

• I’m a visual person so again on a map, show me the good schools, and only houses under lets say $4k/sqm within x minutes from those schools by bicycle.

• Overlays like crime, air pollution etc.

• Allow me to do virtual tours of homes. A badly filmed tour through the house with phone + some app that automates Gaussian splatting would make this possible already.

• Show me all monthly fees/property taxes

• If the house/apartment is mostly empty, rundown or has ugly interior then allow me to take some pictures of my stuff which can be inserted into listing pictures to make it easier to imagine how it looks if I actually lived there.

• Etc

So it seems a vastly cheaper and improved experience can be offered here using AI for both agents and consumers with tech that has already existed for several years. A lot of these platforms are fairly sleepy and have not really done this because they were a monopoly protected by network effects. And I think the market is probably right to dump them recently. Even if they stay on top, there will be pressure.

What would a disruption look like? The idea is that LLMs will act as a portal by aggregating data on listings.

Arguments in favor of this thesis:

• The journey to buy often starts on google/AI bots. You might want to check what good neighbourhoods/schools are etc. So this would be a direct funnel for Chatgpt.

• LLMs would set up a portal for agents who have a strong incentive to upload their listings there as large LLM companies already have a lot of users.

• LLMs would have to build cool AI tools that might entice agents and consumers while portals might be slow to adopt.

• LLMs have data that probably gives info about preferences already based on past conversations.

• LLMs aggregate this and you would have many cheaper rightmoves.

Counter arguments against that are:

• I wonder why Google didn’t do this.

• It would still require a large amount of real estate agents to sign up relatively quickly before Rightmove adapts. Real estate brokerage markets are very fractured. If asking LLMs gives you very incomplete listings, people will still end up going to Rightmove first as they still are by far the largest funnel for people looking to buy.

• Rightmove has a real proprietary data advantage for real estate brokers that would take at least multiple years of high density operation to break:

  • Real time demand/intent signals on a national scale such as searches, saves, alerts, repeat visits, location pivots, budget changes, time-on-market attention

  • Longitudinal listing history

  • Closed-loop performance and benchmarking

• Consumer habits are sticky. Rightmove (and other dominant platforms) probably have the funnel advantage still over LLMs. So there is some room for error for them.

• LLMs would have to set up dedicated operations for real estate with apps with some of above mentioned missing features + what portals already provide. Sales force to get the brokers. It could not be a side thing.

• ChatGPT and maybe Google would be the main contenders as they control most of the consumer market. Openai is bleeding money. Do they have the budget to attack all these niches, with a highly uncertain winner takes all dynamic?

• Specialists > Generalists, so Rightmove still has an edge here IMO. You can see how Google with all its supposed brilliance and its huge user funnel has consistently failed to compete in a wide range of markets. They are almost pathologically unable to build stuff that people want.

• Building + running all this costs $$$ even if AI makes it cheaper, so LLMs will have to charge estate agents at some point. Rightmove has 67% EBIT margins, if LLMs offer it for 50-75% off, their margins would not be that large.

• It would still be a winner take all kind of thing as people are drawn to the place with the most complete listings. So even at 50% off, if you have to pay 3 LLMs to show your listings it is actually more expensive if you are an agent.

• This means it is literally an existential fight for websites like Rightmove. So you are fighting a cornered specialist to the death while your only real modest advantages are the funnel (as models themselves for this task are a commodity basically) and a small data advantage from past chat logs. But overall data advantage would probably be negative.

• What if we end up with 6-7 dominant LLMs? You kind of need a monopoly like funnel to disrupt this. Rightmove doesn’t need a large advantage to catch most of the traffic. If that is the case Rightmove might be another dominant real estate LLM. I suspect we will get all these specialist LLMs over time where the large ones like Openai will still be dominant, but it won’t be like Google.

So I am somewhat skeptical ChatGPT (who would be the only real contender with maybe Gemini) can seriously disrupt this. If they try it will probably hit Rightmoves margins. They might have to lower fees to stay dominant and invest in building new features at the same time.

Read more

Read more

Read more

BUT I will continue writing up new ideas for another 6 months.

Then I will cease writing paywalled articles. I lowered my monthly price to 30 euro’s/month. Anyone subscribing or renewing annual or founding after today (January 12th) will not get their money back! You can do a charge back though if you read this post a few days from now and it renews. But I won’t refund it (I will lose >10% on that due to fees). So if you want to keep reading you can cancel your annual sub and just sub monthly for 30/month until June 2026. I will then pause subs entirely in mid June.

Then whoever still has an annual paid subscription running from before today (the 12th of January 2026) can get their money back for the remainder in July if they wish. After that I will still write the occasional free post.

Edit: Don’t monthly sub if you still have a few months to go on current sub! Otherwise you will be double subscribed.

As for the next 6 months, I suspect software stocks will be the main focus due to the narrative that AI will destroy software. Everyone on Twitter seems very bearish on them. As I am building a tool for myself using AI (Claude Code) to make finding mispriced stocks and special situations more efficient I kind of have a front row seat to this.

It seems the bear thesis has two parts:

• Customers will just build products inhouse.

• SaaS market will be flooded with new competition from smaller companies as AI lowers barriers of entry

I think the first point is completely ridiculous for 99% of SaaS services. While there will be consumers and companies building vibe coded apps for very niche use cases, most SaaS apps such as accounting, EHR, logistics software etc will not be built inhouse as economies of scale kick in very very rapidly for most customers. There will always be weird edge cases and bugs that cost way more in time spent to fix than the 3-400/year you are paying to some well established service that has these ironed out already. Especially when other things such as security are very important.

As for the second point, I think this assumes that the only moat in software is cost of building an app. Assuming that somehow most software was gate kept by a lack of software engineers, prohibitively high costs. I would say mid to large scale software companies primarily get their moat from:

• Stickiness, can be broken up into:

  • Employees having to learn a new work flow (AI wont help much here).

  • Time cost of switching (AI will erode this one)

  • Something going wrong and losing a customer worth 10x the cost savings (AI will help with this one, but will take a long time before perception of this risk goes away).

  • Custom made rules, automations, fields etc. Building this requires feedback with the customer which is a headache, AI won’t help much here.

• Network effects. Everyone usually thinks of facebook here. But for example employees knowing how your software works so they can be more productive at it is kind of a network effect, (although technically this belongs in stickiness basket as well.) These can be local or global. AI does little to erode this.

• Data Moat. AI actually widens this moat. As new entrants will simply lack the data to provide same quality service.

• Integrations with other software. Not easy to achieve for a new entrant. Especially if trust declines when a lot of these vibe coded apps start causing damage.

• Good track record of reliability if that is important (for example very high up time, security or compliance). A new entrant would have to be much better in some quick and obvious way.

• Distribution advantage. When you are the major player, when customers search for a solution that is first one they will find.

• Economies of scale. AI tools won’t mean code becomes free. Larger companies can spend on more security, new features, customer service, lower hosting costs etc.

• Certification requirement. This one is somewhat rare, but I have seen them. Where it is near impossible for a new entrant to break in as government doesn’t like sensitive data spread out over a dozen companies for example.

Finally I have said this before, but the more complex/specialized a piece of software is, the worse these AI tools are at building it. To the point where an experienced dev is more productive just writing it themselves.

I do think in order to win this time around, you need companies that actually generate earnings (SBC is a real expense!) or some other strong catalyst. PE ratios will matter. So buying some cash burning SaaS at 5x sales that sold off 50% from 10x sales might not be a winning strategy anymore.

Also good company culture matters. I think for quite some time these AI tools will work best in the hands of experienced devs. And most experienced devs work in some capacity for software companies. So companies with decent code bases and good talent will likely become significantly more productive.

My prediction is that in a year or two as the strengths and weaknesses of these AI tools become clear people will still remember that a lot of SaaS companies have really nice recurring revenue, solid potential for a moat and need essentially 0 capital to grow revenue once already breaking even and the good ones will end up doing well.

Read more

• Beretta has purchased a 10% stake and seems intent on buying the whole company (a poison pill has been adopted if they go over 10%).

• It seems operational optimizations have been made that will hopefully improve margins and revenue in 2026.

• This is a cyclical stock trading near 16 year low going into likely some very contentious elections in the next 2 years.

Up until recently this company was regularly printing around $100m of EBIT/year but it seems now with Trump as president and a lot of Americans probably owning a lot of guns already sales have slumped:

Market cap is $500m with $80m of net cash and the main reason I am interested is

Read more

Read more

Read more

There are currently no large volume spikes on the HK exchange. Unfortunately the auditor (PWC) resigned early this year about a related party transactions dispute. Which is most likely with Ping An which owns ~75% of Lufax Shares. So the company has delayed filing its annual report and the stock has stopped trading on the HK exchange.

What piqued my interest though were some recent unusual trading patterns on the US options exchange, someone seems to be backing up the truck on options (which are normally quite illiquid as Lufax is a $2-3bn Chinese stock with a small float).

Oct 17 2025 $4 call options:

Nov 21 $4 calls:

Dec 19 $3 calls:

Dec 19 $4 calls:

Mar 20 2026 $2 calls:

So that is several $100k in option trades over a fairly short amount of time. Maybe somebody knows something? Maybe Ping An will buy out the remaining 25% they don’t own? Lufax trades only at about 22% of tangible book value. If Ping An offers 40% that is nearly a double from here. Those $4 options would be well in the money. Ping An has a fairly strong incentive to do this. They did something similar recently with OneConnect (Ping An owned 65% before taking it private at a fairly sizable premium).

Even if this is not some insider backing up the truck on insider info like last time, there are reasons to be bullish:

• Lufax has called an extraordinary general meeting on September 30. One of the things on the agenda is increasing its annual caps relating to Ping An consumer financing business. This seems bullish for demand.

• The Chinese government has set up an interest subsidy on consumer loans.

• CCP gives support to consumer loan companies like Lufax (cheaper financing)

• CCP has introduced a lot of new regulations in the past 1.5 years which will benefit Lufax and hurt its smaller competitors.

• As I have mentioned before, Lufax has switched to a new model where they basically take 100% of potential loan losses up front in a fairly conservative way. This has artificially depressed margins and should no longer be a headwind going into 2026. Earnings could surpass $500 million in 2026. TIKR analyst estimates sit at around $450m for 2026.

• Lufax still has about $1.5-2.35 in net cash/share on their balance sheet (depending on how you count). Vs $3.55 share price right now.

Given all this, I am back in Lufax with a fairly sizable position. I see this more as a trade than a long term investment though. Downside should be fairly limited here, while upside could be substantial.

Disclaimer: Readers of this blog should do their own due diligence before buying or selling any of the mentioned stocks, since I have been wrong before and cannot guarantee all information in this write-up is 100% factual. I may buy or sell the above mentioned stocks at any time. Past success is no guarantee for future success. Some of the stocks mentioned might have poor liquidity, so make sure to check average daily trading volume before buying or selling anything. I am not your financial advisor.

Subscribe now

• Has reduced share count in the past 10 years

• Has a large part of its market cap in net cash

• Started returning cash

• Has a faster growing higher margin segment that is now approaching half of sales

• Trades at single digit PE

• Shows early weak indication of possibly take-private

Read more

Also closing ManpowerGroup China (2180) here after the 1.6 dividend. >50% return depending on when you got in or out. The trade was basically the special dividend, that happened so now I move on.

Planning to completely overhaul the blog in the next 6 months or so. Will be a completely new website outside of substack. With a much higher quantity of (hopefully still quality) ideas. Will probably write a bit less as I am working on this, but paid subscribers will get free first year preview to make up for it. I don’t want to say too much about it now, but I am pretty excited about it.

Subscribe now

• German tanks are shooting at Russian tanks in Kursk

• Xi Jinping, the Chinese communist leader, is egged on by investors to create more social stimulus and social safety nets but he hesitates as he thinks it will make Chinese people weak and complacent.

• Bitcoin is at $100k

• Donald Trump wins a second term after somehow taking blue collar and middle class voters away from the Democrats after going on the Joe Rogan podcast saying he wants to be a whale psychiatrist after he is done being president.

• Democrats campaigning with Dick Cheney against Trump.

• In Syria a suit-wearing ex-ISIS/Al Qaeda jihadist defeated Assad in under 2 weeks and seems obsessed with building institutions, vaguely respecting minority religious groups and not fighting the West. Shakes hands with Trump, Republicans seem to have no problem with this.

• Interest rates at 4.5%, the S&P 500 trades at 30x earnings with record high profit margins while Trump is doing smoot hawley 2.0

• The US government makes a $9 billion investment in Intel taking a 10% ownership stake.

• AI that can communicate like humans and solve many problems better than most humans.

• Warren Buffett converts Berkshire’s cash reserves into Bitcoin saying “now that sour puss Munger is gone, I can finally do what I really want! CRYPTO 69 420 BITCHES” He is seen at the Berkshire annual meeting wearing a baseball cap backwards.

I wonder if they can guess which one of those is not true. And I wonder what a list like this would look like 10 years from now that would seem unbelievable to people now. Aliens making contact with humanity? America adopting Communism (led by Republicans)? You really have to start reaching for truly crazy things to top this right?