I Sold Everything in One Hour

Tuesday. 11am.

I opened my brokerage, looked at a portfolio I'd held for years, and started selling. VOO. QQQ. SCHD. UNH. All of it.

By noon I'd rebuilt the entire portfolio from scratch. The portfolio I woke up with looked nothing like the portfolio I went to bed with.

It took one hour.

It took me six months to get here.

I'm publishing this with a date on it because I want a record. Not of the trades — of the thinking. So that in five or ten years, anyone can come back to this page and see exactly what I believed, exactly when I believed it, and whether I was right or wrong.

This is a timestamped bet. Here it is.

What You're Actually Holding

If you own the S&P 500 — and statistically, most of you do — let me tell you what's actually in it. Because I don't think you've looked. I hadn't.

34-41% of VOO and QQQ is seven companies. Apple, Microsoft, Nvidia, Amazon, Google, Meta, Tesla. You're paying index fees to hold a concentrated tech bet you didn't choose.

But here's what should keep you up tonight.

Buried in those indexes — in the other 493 stocks — is a graveyard of business models that are dying right now. SaaS companies whose moats are evaporating.

Financial services firms built on human intelligence processing. Insurance companies built on information arbitrage. Consulting firms that charge for expertise AI makes free. Legacy corporations still priced as if human knowledge work will remain scarce forever.

15-20% of the S&P 500 is companies whose core business model is the intelligence premium — charging you because someone there knows something you don't.

That premium is going to zero. Not slowly. Not over decades. Over years. And when it does, those companies either adapt or die, and your index holds them the whole way down while the weighting formula catches up.

That's not diversification. That's dilution. You're watering down your winners with companies you would never buy individually if you actually thought about what they do and whether that thing survives the next ten years.

I'm writing this in February 2026. Come back in 2030. Look at which companies in the S&P 500 lost 50-80% of their value. Then check whether they were in your index fund the whole time.

They were. You just didn't know.

The Hypocrisy That Broke Me

I'd been writing essays about all of this for weeks.

The intelligence premium is dead. Compounding might be broken. The structural relationship between corporate profits and consumer spending is being rewired by AI. Labor share at historic lows and falling. The economy splitting between companies that own AI infrastructure and companies that get consumed by it.

I believed every word.

And my portfolio was 57% in VOO and QQQ.

I was funding the old economy while writing essays about why it's dying.

That's a special kind of dishonesty. Not the kind where you lie to other people. The kind where you lie to yourself. Where you maintain two separate realities — the one you write about and the one your money lives in — and you never force them to meet.

Belief without action isn't conviction. It's commentary.

I'd been a commentator for six months. Seeing clearly. Writing clearly. And holding a portfolio that contradicted everything I said.

Here's what holding felt like from the inside. Every morning I'd see VOO. Think: "This doesn't fit my worldview anymore." And do nothing. Because selling feels permanent. Because holding feels safe. Because cognitive dissonance is comfortable — you can believe one thing and do another for a very long time as long as you don't force yourself to choose.

Holding is a decision. It's just a decision you make by doing nothing, which means you never have to feel the weight of it.

I think most of you are doing this right now. And I think you know it.

What I Believe

Here's the thesis. Plainly. On the record. Dated February 26, 2026. So there's no ambiguity later about what I was saying and when.

The economy runs on a stack.

At the bottom is energy. Above that is compute infrastructure — chips, data centers, networking. Above that are the application layers — companies that turn infrastructure into products people use. Everything above that — the knowledge work, the information processing, the human intelligence premium — is getting compressed toward zero.

Most portfolios are overweight the top of the stack. The part that's dying. And underweight the bottom. The part that everything else depends on.

Energy is the binding constraint for the next decade.

AI compute is projected to double data center power demand by 2030. Microsoft restarted Three Mile Island specifically for AI. That's not a signal. That's a declaration.

Nuclear is the only baseload source that scales without carbon constraints. Uranium supply is constrained from a decade of post-Fukushima underinvestment. The companies that own nuclear capacity today are AI infrastructure companies that the market still prices like sleepy utilities.

I'm calling it now: by 2030, energy companies will be recognized as the backbone of the AI economy the same way semiconductor companies are today. The rerating hasn't happened yet. It will.

The intelligence premium is dying.

Every company whose model is "we know something you don't" is on borrowed time. Insurance brokers. SaaS companies. Consulting firms. Financial advisors. Legal services.

Three weeks in February 2026, I watched the market start pricing this in — $2 trillion wiped from software, cybersecurity cratered, insurance hit, trucking hit. That was the opening act. The repricing will continue for years.

Labor share keeps falling.

53.8% and heading toward 45% by 2035. The economy grows but growth goes to capital and compute owners. Consumer spending — 70% of GDP — runs on wages.

When wages compress, spending compresses. The consumer economy compresses. And the index that tracks the consumer economy compresses with it.

The broad index is a bet on the old economy continuing.

This is what people don't see. VOO isn't neutral. It's a bet that the current distribution of economic value persists. That the companies in the index today will capture value the same way ten years from now.

If the structural relationship between labor and capital is changing — and it is — then the index carries dead weight you're paying to hold.

That's what I believe. Every word of it. On the record.

What I Bought

Everything fits into three layers of the stack.

Energy

Constellation Energy. Largest nuclear fleet in the US. Microsoft signed a 20-year power purchase agreement for AI data centers. Priced like a utility. It's not.

Vistra Energy. Nuclear plus natural gas. Same thesis, more diversified, cheaper. Best value in the portfolio.

Cameco. Largest publicly traded uranium producer. Nuclear is the energy layer. Someone has to mine the fuel. Supply constrained. Demand accelerating.

Infrastructure

TSMC. Manufactures every advanced AI chip on Earth. Every fabless company — Nvidia, AMD, Apple, Broadcom — depends on this one company. Geopolitical risk is real. The monopoly is realer.

Broadcom. Custom silicon for Google and Meta. AI networking. Every dollar that shifts from GPU rental to custom chips flows through Broadcom.

Equinix. Largest data center REIT. AI inference happens somewhere physical. Equinix owns the somewhere.

Application

Google. Cloud, custom chips, YouTube, DeepMind, Waymo. 87% of revenue is advertising — the tax on human attention. Attention doesn't get disrupted by AI. It gets amplified.

Brookfield. Added to existing position. Infrastructure, renewables, data centers. A trillion in assets positioned where the world is going.

Berkshire. Trimmed but still core. New CEO immediately deployed into Google. A cash fortress run by an infrastructure operator. More thesis-aligned than people realize.

What Could Go Wrong

I'm 27 years old writing about my own money. I could be dead wrong.

If AI progress stalls, broad indexes win. If energy demand projections miss, the nuclear thesis is overpriced. If custom silicon doesn't erode Nvidia's moat, Broadcom is expensive. If the consumer economy holds up better than I think, the labor share thesis is premature.

I'm not pretending certainty. I'm taking a position.

The downside is I underperform the index for 3-5 years.

The upside is these companies compound for 30 years as AI infrastructure becomes the defining economic buildout of the century.

I can live with the downside. I couldn't live with the hypocrisy.

The Record

I want to say something to the person reading this in 2030. Or 2035.

You're going to know things I don't know right now. You're going to know whether the nuclear thesis played out. Whether TSMC survived the geopolitical risk.

Whether the intelligence premium really went to zero or just compressed. Whether Constellation and Vistra rerated the way I think they will. Whether Broadcom mattered or Nvidia kept its moat.

You'll have the answers. I only have the reasoning.

But here's what I want you to notice.

In February 2026, most people held the index. Most people trusted the compounding gospel. Most people had portfolios built on assumptions from 2019 or 2021 — assumptions about which industries grow, which business models work, what the economy rewards.

And most people never revisited those assumptions, even as the world changed underneath them.

I'm not saying I'm right. I'm saying I looked. I examined what I owned. I asked whether it still made sense in the world that was forming. And when the answer was no, I acted.

Most people didn't. Not because they're stupid. Because holding is comfortable and selling is scary and cognitive dissonance is the easiest thing in the world to maintain.

If I was right — if the energy thesis played out, if the intelligence premium really died, if the index carried dead weight all the way down — then this page is a record of someone who saw it and moved.

If I was wrong — if VOO outperformed everything I bought for the next decade — then this page is a record of someone who had the courage to be wrong in public with a date on it.

Either way, it's honest.

And that's more than most people's portfolios can say.

The Question for You

I'm not going to tell you what to buy. I don't know your situation. I'm not a financial advisor. This is a 27-year-old's thesis, not gospel.

But I am going to ask you one question.

When was the last time you opened your portfolio and actually looked at what's inside? Not the balance. Not the performance chart. The actual companies. The actual business models. The actual bet you're making about what the economy rewards in 2030 and 2035.

Because if your portfolio is a bet you made in 2019 — if it's built on the assumption that human intelligence stays scarce, that SaaS moats hold, that the consumer economy runs on the same engine, that the index you picked will contain the same winners a decade from now — then you're not investing.

You're holding a thesis you haven't examined in years.

And holding a thesis you haven't examined is the most expensive form of doing nothing there is.

I don't know what your version of Tuesday looks like. But I think it's coming.

When it does, I hope you're ready. And if you're not — if you read this in 2026 and did nothing, and you're reading it again in 2030 wishing you had — at least you'll know that someone told you.

With a date on it.