Where engineering margin becomes financial margin.

Every engineered system runs with margin — headroom between what it does today and what it physically cannot exceed. When that margin runs out, roadmaps slip, capex gets rewritten, and the demo that went viral quietly never ships. This site is about finding those limits before the market does.

I'm a robotics software engineer working in semiconductor manufacturing. I write about semiconductors, AI infrastructure, and robotics from the perspective of someone who ships this work, using only public sources: earnings calls, patents, papers, and datasheets.

Three things I write about

• The physical layer. Fabs, process nodes, packaging, HBM, equipment — why only a handful of companies can make any of it.
• The bottleneck watch. Power, cooling, materials, capacity. What is actually constraining AI buildouts right now, and what breaks next.
• Physical AI, reality-checked. What a robotics demo does and does not prove, and what would have to become true for it to ship as a product.

All three end in the same place: what it means for margins, capex, and the companies you may own.

This site is independent. Nothing here is investment advice, and views are my own — not those of my employer.