Updated 26 September 2026

Huliwood.

I write about what I can’t stop thinking about
and invest in the ideas I think survive.

I’m Julián Martin. Half German, half Spanish, 100% insufferable. Kidding. Mostly. Further down, one of those ideas gets scored in public.

Latest

Essay · 20 September 2026 · 375 words

Dog-Sitting a Life

Walking Someone Else’s Arrangement Read it

Two that aren’t the latest: The Last Leash (why techno-feudalism might be the good ending) and You Can Have Everything (which I reread whenever I catch myself trying to).

No schedule, no filler. Get the next one by email.

One idea, all the way round

How I work: think, write, spot change, map to companies, returns, freedom, and round again. Here is one idea taken all the way round, The AI Landlords, with the score it has earned since.

Think

“Once a century, the economy changes workers.”

The first line. Two sections on, the essay sets out the pattern.
  1. 1850s: Coal miners needed housing near coal seams. Mine owners got rich.

  2. 1950s: Accountants needed offices near city centers. Skyscraper tycoons got rich.

  3. 2025: AI agents need data centers near cheap electricity. Blank gets rich.

Fill in the blank.

Write

Essay · 14 Dec 2025 · 1,815 words

The AI Landlords Manhattan for Compute

“Bitcoin miners became AI landlords.”

Spot change

“Most people saw ‘cloud computing.’ I saw the real estate play nobody recognizes as real estate – except the tenants aren’t humans.”

The chart that started it: “for the first time in history, data center construction is overtaking office construction in the US.”

Jan 2014 to Jul 2026

Data centers passed offices in November 2025 and have stayed above every month since. From the essay’s month to July 2026 (preliminary): data centers +35.5% to $75.2bn a year, offices −6.7% to $48.2bn. Redrawn from US Census data, seasonally adjusted; “offices” is the office total less data centers.

Map to companies

“…a thesis that started with a chart and ended with 44% of my portfolio in companies most people have never heard of.”

Named in the piece, 14 Dec 2025

IREN and CIFR (the 44%), TSLA and PLTR, with QQQ, the Nasdaq 100, as “my ‘I might be totally wrong’ hedge”.

Where they are, one site each: IREN in Childress, Cipher in Odessa and Tesla in Austin, three of the four on the Texas grid, near the wind farms west of Abilene. Palantir writes software and lives in Denver.

IRENCIFRTSLAPLTR

Returns

−18.3 points vs the Nasdaq 100

The four, equal-weighted: +4.1%. QQQ, the Nasdaq 100: +22.4%.

IREN +21.0% CIFR +12.0% TSLA −17.4% PLTR +0.8%

12 Dec 2025 to 22 Sep 2026

Why I might be wrong, as I wrote it

“The bear case is simple: efficiency gains outpace demand growth.”

Scored from the last close before the piece ran, equal-weighted, dividends in, refreshed weekly. It scores what the piece named, not what I hold today. None of it is advice.

The whole book by year, % in euros

Return2020*202120222023202420252026†
Me+87+55−66+117+89−14+31
QQQ+31+38−28+49+34+7+24

Since April 2020: +362% against +243%. Since January 2021: +147% against +163%.

Time-weighted, every account I hold. Years rounded; the totals use the exact figures. *From 20 April. †To 24 September.

Freedom

Think → write → spot change → map to companies → returns → freedom

Started investing in April 2020, two days before I turned 26; started writing three Aprils later. Sold everything I owned in 2024 and left with a backpack. What I tell people I do:

Human language model. Content cringeator. Funded by the lunar reptilians who control the tides of finance.

Currently hopping through AUS 🦘

Your turn