

The global hardware shortage has only just begun. 🤖 Very few people are currently using AI properly—most only occasionally ask light questions using free accounts. As Musk said, a significant portion of future memory will go not to data centers, but to physical AI—robots—because robots are ultimately what will truly enhance human productivity. When robots take over not just factories but household chores too, the demand for memory and compute will far exceed what we can imagine. Building a single memory fab takes years. Regardless of debates around the AI bubble, supply shortages won’t disappear easily. And powering that compute ultimately comes down to electricity and cooling. (Not investment advice)

Bitcoin is the greatest asymmetric opportunity I’ve ever encountered. ₿ In a world where AI can generate almost everything infinitely. An asset that never increases in supply. Future generations will look back and say: “Bitcoin was that cheap back then?” And ultimately, securing the network comes down to electricity and cooling. (Not investment advice)

ECB President Lagarde identified hyper-scalers’ bond issuance as one of the reasons for rising long-term interest rates. Insurance companies and pension funds have traditionally been the buyers of 30- and 40-year bonds. Their allocation to U.S. Treasuries has fallen to the lowest level since the 1950s (BofA). Yet, both the U.S. Treasury and Big Tech are simultaneously flooding the market with ultra-long-term bonds targeting the same investors. Data centers are ultimately built with debt. Transformers, cooling systems, and buildings are decades-long assets requiring decades-long funding. As interest rates rise, construction becomes more difficult, and those who can secure cheaper electricity and cooling gain the advantage. At its core, AI is ultimately a problem of electricity, cooling, and capital.

The battleground for AI companies has shifted. Anthropic has already built the best software business. But the fight ahead isn’t about “who can build a smarter model.” It’s about who can secure the chips and electricity faster and more cheaply. Code can be copied. Substations and cooling infrastructure cannot. In the end, AI is a problem of electricity and cooling.

Are people selling AI stocks to buy AI stocks? There’s talk in the market that when a mega AI IPO like Anthropic approaches, funds may sell off some of their existing big tech and semiconductor stocks to raise capital for the new offering. If true, the money isn’t leaving AI—it’s just shifting within the AI ecosystem, from hardware to model companies. But one thing remains unchanged on the ground: No matter which model wins, it still consumes electricity and generates heat. Funds may move around, but demand for power and cooling does not.

Binance hacked for $6.3 billion—we underestimated North Korea 🚨 How can a country with electricity shortages produce the world’s top hackers? The answer: asymmetric warfare. A single missile requires steel and foreign currency, but one hacker needs only a single room and a few computers. Success brings in billions of dollars. North Korean hackers are known to operate overseas while concealing their identities. By day, they work as remote developers for Western companies and Web3 projects; by night, they exploit vulnerabilities they’ve observed within those systems. The most dangerous weapon isn’t code—it’s people. A fake high-paying job offer on LinkedIn, a single malicious file. This time, it wasn’t a wallet that was breached—it was a developer’s laptop. For a country cut off from SWIFT, cryptocurrency is a lifeline beyond sanctions. The lesson is simple: 1. Never open job offer files from unknown sources. 2. Separate your work computer from your personal wallet. 3. Store coins you plan to hold long-term in a personal wallet, not an exchange. No matter how large the exchange, if one person is compromised, it’s over.


