Capex Signal is a weekly note on AI infrastructure spending and the practical implications for semis. I'm starting where signals are cleanest: semiconductor equipment (WFE / semicap). Over time, I'll expand outward to AI chips and memory/HBM, because that's where capex becomes revenue.
The Core Idea
In this sector, the market gets loud fast. To stay grounded, I use a simple hierarchy:
- 1.Constraint (what is actually limiting output?)
- 2.Intensity (how much more equipment/process complexity is required?)
- 3.Durability (is it a one-quarter timing issue or a multi-year slope?)
Most "AI capex" arguments break because they skip step (1).
What You'll Get Each Week
Each weekly post has two layers:
Free Layer (investor-friendly):
- •8–12 bullets: what changed + why it matters
- •Short "so what": how I'd interpret it and what to watch next
Premium Layer (technical depth):
- •The mechanism in plain language (process/physics intuition, not jargon for its own sake)
- •A durability check: base/bull/bear framing
- •A short "what would change my mind" trigger list
- •Sources / receipts
- •A small options/trading education corner (general, risk-defined; not personalized advice)
For now, I'll publish primarily free while the format gels. If the premium layer proves useful and readers ask for it, I'll turn on paid access gradually.
How to Use This
If you're a long-term investor:
treat this as a framework and a watchlist of triggers.
If you trade around earnings:
treat it as a way to separate real cycle changes from one-off sentiment moves.
Scope + Rules
- •Focus: semicap + semis (AMAT/LRCX/KLAC/ASML + NVDA/TSMC/AMD + memory/HBM) and selected AI apps (e.g., PLTR).
- •No crypto.
- •Not investment advice.
If you want a starting point, begin with the most recent weekly post, then use the archive to follow themes like AI capex, packaging intensity, and memory/HBM constraints.