Explainers
Concepts, decoded. Plain-English breakdowns of the ideas driving the market β each one linking to the stocks it moves.
Take-or-Pay Contracts Rewrite Who Owns AI Demand Risk
Take-or-pay contracts require a buyer to pay for reserved compute whether or not it uses it. That structure lets AI infrastructure sellers borrow against contra...
Hyperscalers Are Spending 750 Billion Dollars on AI. The Bill Comes Later.
Capex is cash out the door today; depreciation spreads the pain over years. That gap is the real story of the AI buildout: 750 billion dollars planned for 2026,...
The Network, Explained: AI's Third Bottleneck Hides Between the Chips
The AI buildout's first two bottlenecks β memory and power β get all the headlines. The third is the network between the chips, and it decides how much of that ...
Don't Pay the Toll: RISC-V vs. ARM and the War Over AI's Brain
Forget the GPU battle for a second: the quietest war in AI is between ARM, whose designs sit inside virtually every smartphone and collect a royalty on each, an...
Space Launch Economics, Explained: How Reusability Broke the Cost of Reaching Orbit
Reusable rockets turned space from a government-only luxury into a commercial industry. Here's how the economics work, who's winning, and what cheap launch unlo...
Power, Explained: The Real Reason Your AI Data Center Can't Plug In
The most expensive part of an AI data center isn't the chips anymore. It's getting someone to turn the lights on. Front-of-the-meter vs behind-the-meter, and wh...
HBM, Explained: Why the Memory Stack Next to the GPU Is AI's Hottest Choke Point
The AI buildout has a second bottleneck nobody talks about: the memory stacked next to the GPU. This explainer breaks down what HBM is, why it's so hard to make...