E-commerce operator and investor. Tracking the AI memory super-cycle and the companies powering it.
Jun 25, 2026
•
5 min read
Micron just printed a record breaking quarter. That’s exactly why investors shouldn’t hear only the bull case. Here’s the strongest argument against my view.
Jun 17, 2026
6 min read
Jun 15, 2026
4 min read
You own one phone. Soon you'll run a hundred AI agents. That's the whole difference.
Jun 11, 2026
Memory isn't trending right now. It's swinging with big energy.
Jun 5, 2026
3 min read
Three things hit at once, the memory theme stock dropped hard, and none of them broke the thesis. Here is what actually happened.
Jun 1, 2026
I am up around 60% on my DRAM position. Everyone says take some off the table. Here is why I am not.
May 28, 2026
2 min read
Every major player in the semiconductor supply chain confirmed the same thing this week from a different direction. Here is what happened.
May 26, 2026
Every time someone says AI efficiency will kill memory demand, I think about human resources and my payroll
May 21, 2026
I watched the Samsung strike saga play out in real time from Hong Kong. Checking Yonhap at midnight. Watching Korean stocks move before New York woke up. What I saw wasn't just a labor dispute. It was a masterclass in why corporate structure matters as much as technology.
May 19, 2026
Three major memory stories dropped today. All three confirm the supercycle is real. One of them deserves more attention than the others.
May 18, 2026
1 min read
Three things happened in Korea today before US markets even opened.
May 14, 2026
I've been running an agentic AI on my e-commerce business for months. Here's what it's taught me about where memory demand is actually going.
May 13, 2026
The most common pushback I get on my DRAM thesis is China. "Won't they just flood the market and kill pricing?" Here's why that argument doesn't survive contact with reality.
May 12, 2026
A Korean policy adviser posted something on social media this morning. Foreign investors sold $3.8 billion of Korean stocks within minutes. Here's what actually happened, and why I didn't sell.
May 11, 2026
I'm long DRAM. I've already written three posts laying out why I think the AI memory supercycle is real and has years to run. Now I'm going to make the strongest case against my own position.
May 9, 2026
Reuters reported something so absurd yesterday that I genuinely had to read it twice and assess if it is fake news.
May 8, 2026
Samsung's chip minions rejected a $340,000 bonus this week, because it was offered as a one-time payment, they know what's up and the supercycle is not ending anytime soon.
May 7, 2026
The bottleneck isn't compute. It's memory. And almost nobody is positioned for what comes next after agentic AI takes over industries.