(#145) 💿 Nvidia is 'all in'; 🚙 Who’s doing the strategy at BMW?
A parallel between 🍎 Apple and 🔴 Huawei
Thank you for being one of the +4,500 minds reading this newsletter
Here is what you’ll find in this edition:
[Essay] Apple’s lost decade
An Eastern European Saga
A parallel between 🍎 Apple and 🔴 Huawei
NVIDIA is “ALL IN”
📈 My investment thesis (part 1)
…and more 👇
Onto the update:
[Essay] Apple’s lost decade
Apple made a strategic choice a decade ago and it chose comfort (and a lot of money)
Instead of building the future, it took a dividend from the present. While Google plowed billions into search, data infrastructure, and foundational AI research, Apple signed a check to license Google Search as the default on Safari. That check, estimated at over $20 billion per year, was enough to keep Apple’s services revenue growing, but it was also the price of a decade-long delay in understanding how to work with data at scale.
Today, that decision has metastasized into the embarrassment that is Siri, the me-too nature of “Apple Intelligence”, and the company’s palpable absence from every major AI inflection point.
...Continue reading HERE.
An Eastern European Saga
Putin keeps insisting his war is about NATO and “nazis”, but the more boring and therefore more accurate explanation is that it’s about GDP math (ie. prosperity). Since 1990, the former communist countries that joined the EU have grown their economies almost 10x, while Russia has managed barely 4x.
🇵🇱 Poland, Slovakia, the Baltics, 🇷🇴 Romania: skyscrapers, highways, and EU per-capita income curves that leave Russia in the dust. And sitting right next door is Ukraine - the “tipping point state”. If 🇺🇦 Ukraine prospers as an EU member, Russians will start asking awkward questions: “Why do people who look, speak, and eat like us have a functioning economy while we have crony oil rents and kleptocracy?”
That’s why the war isn’t really foreign policy, but it’s internal policy disguised as external aggression. Ukraine can’t be allowed to thrive, not because of NATO tanks but because of Russian pensioners with Youtube and cousins in Warsaw. The invasion is basically regime maintenance: stop the mirror from reflecting back to your own people how badly you’ve run the place.
If Ukraine’s equity story is “Eastern Europe at EU multiples”, then Russia’s governance risk is that shareholders (citizens) realize they’ve been holding the wrong stock all along. WSJ
A parallel between Apple and Huawei
Two mirror books: “House of Huawei” (Eva Dou) and “Apple in China” (Patrick McGee). Read together, they show how elite firms are bent and turbocharged by China’s political economy.
It was a coincidence that I read them one after another, so I got the chance to observe some similarities between these 2 companies:
1/ work as devotion - both books detail the “best place to work” myth cultures that ask for your all (time, health, identity) in exchange for belonging and velocity. That bargain looks heroic in brochures and costly in real life.
2/ state-enabled scale - Huawei grows as a national champion; Apple’s miracle rides on local (Chinese) government incentives, bonded zones, and a supplier state (hello, Foxconn). Different relationships, but same outcome: public power underwriting private dominance.
3/ control and captivity - described by company towns and closed ecosystems (campuses, dorms, secrecy) create astonishing throughput and strategic traps. Huawei is bound to the party-state, while Apple is bound to China. Each is “captured” by its own success.
If you build, invest, or teach strategy, read both books side-by-side. Amazon
NVIDIA is “ALL IN”
Here’s Nvidia’s $100 billion idea: build “AI factories”. Translation: “really big data centers with a shocking electricity bill”. The logic is circular but elegant: Nvidia gives OpenAI the money to buy Nvidia’s chips so that OpenAI can rent those chips back from Nvidia to make ChatGPT spit out slightly wittier answers. The investment “secures” demand while pretending it’s diversification, which is a very Nvidia thing to do. By investing in NVIDIA, you invest indirectly in OpenAI
The fun part is that everyone knows the economics don’t really add up. Bain says the industry would need $2 trillion in revenue by 2030 just to sustain the capital spending required, and we’re currently short by about $800 billion. But Nvidia doesn’t care. Jensen Huang is running the Microsoft–Intel playbook rolled into one, except instead of Windows on your desktop you’ve got CUDA on your datacenter. It’s less about selling GPUs and more about keeping every ambitious AI lab hopelessly addicted. The factories are just the crack houses. An AI bubble is forming. FT
Who’s doing the strategy at BMW?
BMW has decided that the future of the X5 is… all the futures, all at once:
Petrol
Diesel
Hybrid
Electric
and now hydrogen. 😅
Which is kind of like hedging by buying every option on the board, then bragging about your “technology leadership” because you’re paying five times the R&D, five times the production complexity, and five times the supply chain headaches. Hydrogen, of course, is the funniest line item: there’s no infrastructure, no market, and no realistic economics, but BMW swears it has “an essential role in global decarbonisation”. Translation: “We spent billions on this already and can’t admit it was a bad idea.”
Imagine proudly announcing that your genius strategy is to make cars for every possible energy source, including one that requires filling stations which don’t exist and never will because they’re too expensive. It’s like saying, “Our leadership is so strong we can lose money five different ways”. Meanwhile, Tesla and BYD are doing the boring thing: picking one lane, scaling it, and crushing costs. BMW, on the other hand, is out here building a hydrogen SUV for a hydrogen world that exists only in... powerpoint slides? Leadership indeed. LINK
🇩🇪 Germany is changing too
Germany’s industrial policy for decades was basically: “We make cars. The world buys cars. End of story / policy”. And for a while, it worked. Autos & Parts sat comfortably at 18–20% of the German market, carrying the DAX on their Teutonic shoulders. But the chart says it all: now autos are below 10%, and tech, of all things, is the sector bailing them out. In Germany! The place where “software” was historically something you installed off a CD-ROM in 2004…or a floppy disk 💾
It’s a remarkable reversal: instead of Audi and BMW dictating market weight, you’ve got SAP and a handful of industrial tech firms suddenly doing the heavy lifting. The German car industry spent the last decade lobbying Brussels to regulate tailpipe emissions into fairy dust, while China quietly built an EV juggernaut. Now Volkswagen can’t sell EVs (and is bad at software - see the partnership with Rivian), Mercedes & BMW are playing with hydrogen fantasy projects (ie. building another monopoly), and the market is saying: “You know what, maybe the future isn’t German exhaust pipes, maybe it’s cloud software”. When the DAX looks more like Nasdaq than Detroit, you know the old order has cracked.
No rare earth minerals, no moat
Here’s Europe’s big strategy on rare earths: fly to Beijing, smile for the cameras with Xi, announce an “upgraded dialogue on export controls”, and then fly home hoping the paperwork gets easier. Two months later, 140 licence applications are stuck in limbo, only a quarter resolved, and European manufacturers are bleeding money waiting for magnets and metals that China controls from mine to refinery. This is exactly a REAL preview of what happens when your entire industrial base runs on inputs owned by your geopolitical rival.
What did they think was going to happen? That China would politely keep feeding the EU raw materials while the US and Europe throw tariffs at its exports? If you don’t secure your own supply chain (by investing in EU processing, by building alliances with Africa, Latin America, or Australia) then you don’t really have a supply chain, you have a dependency. And when that dependency moves from rare earths to chips, to AI, to defense tech, the delays will redefine power.
This is the first serious warning and Europe’s response so far is to ask for more “dialogue”. Good luck with that. FT
Premier League and the rest
The top 10 football clubs by revenue are more a business case study in how strategy, branding, and structural advantages compound than a sports ranking.
Real Madrid and Manchester City illustrate Porter’s “differentiation + scale” model where star players and global trophies translate into premium commercial contracts, with broadcast rights amplifying reach.
PSG and Barcelona run closer to “big bets on talent as strategy”, banking on superstars to drive shirt sales and sponsorships, which Rumelt would probably call a bad strategy if it lacks a coherent financial foundation.
Manchester United shows Bradley et al.’s “hockey stick” fallacy, betting on the brand curve to bend upward without making the bold moves (on-pitch results, stadium reinvention) that sustain revenue.
Bayern and Arsenal embody pricing power in Simon-Kucher’s terms. Consistent Champions League appearances and loyal fan bases mean they can extract more value per seat, per jersey, per sponsor.
Two more ideas:
1/ First, football clubs are global entertainment platforms.
2/ Second, the gap between winners and laggards is about structural advantages in commercial deals, market size, and organizational execution. As Porter would put it, the “value chain” for a football club is more about how effectively you monetize millions of eyeballs. Bloomberg
📈 My investment thesis (part 1)
1/ Time arbitrage. I invest long-term (i.e., minimum 2–3 years) so that fundamentals can outperform headlines.
2/ Concentrate to compound. To beat the market, hold max 10 positions. Know them better than anyone.
3/ Indexes are fine. Exceptional businesses are finer. Buffett didn’t build Berkshire by hugging the S&P.
4/ Reinvestment > dividends. Companies that share dividends or engage in buybacks often lack a clear understanding of how to utilize the money.
5/ Moat or bust. I only buy durable advantages: switching costs, network effects, scale efficiencies, brand, regulation, IP, privileged distribution, or proprietary data.
(part 2 to continue)
P.S. In October, I will launch a newsletter on company analysis showing where the moat lies and what my exact portfolio 😎. Of course, under a paywall 🙃