Power the agent.
Nuclear. Agentic infrastructure. Security.
Three eras. One account.
Era 1 and Era 2 are the published record through 17 July 2026. Era 3 is the portfolio allocation of the names.
Era 1 — Pre-framework
+54.4% (Jan 1 – Apr 24)
26 winning tickers / 52 closed lots
Pre-framework, unaudited.
Era 2 — Framework
+16.92% · vs SPX +4.08% · +12.8 pts excess
12 weeks · ~67% cash held · deployed capital ≈ +54%
Era 3 — From 14 November 2026
A one-year catalyst book built around contracted power, full-stack compute, enterprise AI with a reported return, space and connectivity, and the controls agents need. Weights are destination allocations, not current holdings.
Ten equities, then a separate gold bucket.
Three buckets. One portfolio.
US equities, gold and cash, as shares of one portfolio. Gold and cash sit outside the ten.
US equities.
The ten names as a portfolio, then each name. The second percentage is that name’s share of the equities bucket.
| Category | Names | % of total | Why |
|---|
| Name | Role | % of total | % of equities |
|---|
RBRK takes the 7% that was NOW. VST is the path into TLN, not a second power seat.
Each name.
Role, why it is in the book, the portfolio allocation, and what has to stay true.
Why gold.
Gold is the hedge against the capex cycle this book is otherwise long. Direct gold is the monetary holding. Franco-Nevada is the royalty. Southern Cross is a small resource option. They can still fall with the stock market.
| Holding | Role | % of total | % of gold |
|---|
The gold names.
Same card. Direct metal is not the same risk as a miner, a royalty or a discovery.
Research sources · 14 November 2026
Allocation weights are Mute Labs judgments. Issuer figures are attributed to the company. This page is not a live market feed. Destination weights only — no share counts, balances or inferred fills. Prior names kept in retained research are not current destinations.