To The Moon

Mathematical Money | September 27, 2026

Let me give you the quarter first, without the ticker attached, and you tell me what you'd expect the stock to do.

Revenue of $29.6 billion, up 86% year on year. AI revenue alone at $16.7 billion, up 221%, which makes AI more than half the entire company now. Management guiding next quarter to $34.8 billion and saying $350 billion of shipments are already locked in across FY27 and FY28.

That's Broadcom's September quarter. By almost any measure it's one of the better quarters a large company has printed this year.

The stock closed Friday at $352.81. Over the past month it has gone essentially nowhere — down less than one percent. It's also about 27% below where it traded at the start of June, so the damage was done earlier and it has been grinding sideways since.

I bought four September 2027 $310 calls between the 8th and the 10th of September, at an average of $100.08. They're $86.95 now.

Here's the part worth sitting with.

Since I bought, the stock is down about 3.5%. My calls are down 13%.

The stock barely moved. I lost roughly four times as much.

That isn't bad luck, it's arithmetic

Broadcom trades at $353. My $310 call is worth $87. That buys me exposure to a hundred shares — thirty-five thousand dollars of stock — for about nine thousand dollars of premium.

Call it four to one.

That's the appeal, obviously. Same exposure, a quarter of the capital, and the remaining twenty-six thousand stays in the account doing something else.

But a deep in-the-money call moves close to dollar-for-dollar with the stock. So when Broadcom drops a dollar, my option drops roughly a dollar too. Except a dollar is 0.3% of the share price and about 1.2% of my option.

Same dollars. Four times the percentage.

Run it forward: the stock fell 3.5% since I bought, four to one gives you roughly 14%, and the actual move was 13%. The model explains almost the whole thing. Nothing went wrong and nobody made a mistake — that is simply what four-to-one does when the direction goes against you.

Anyone who tells you long-dated calls are just cheaper shares has skipped this paragraph.

Why the quarter isn't being rewarded

Guidance. Specifically $34.8 billion for the next quarter against a consensus sitting near $35.2 billion.

A miss of roughly 1%.

A company growing AI revenue at 221% guided one percent light, and the market has spent the weeks since refusing to pay up for it. If you want an illustration of what "priced in" actually means, that's it. The business is doing brilliantly. It just isn't doing quite as brilliantly as the price had already agreed it would.

This is the thing I keep having to relearn. You don't get paid for a company being good. You get paid for it being better than what the price already says. Those are completely different bets and the second one is much harder.

The risk I'm actually watching

It isn't the guidance miss. One percent on a quarterly number is noise and I'd be embarrassed to sell a 2027 position over it.

It's customer concentration.

Broadcom's top five customers are now 55% of quarterly revenue. A year ago that was 40%. So five buyers decide what this company earns, and they're all building the same thing at the same time off the same capex cycle.

If one of them pauses — not cancels, just pauses — the revenue line moves in a way no amount of design wins fixes quickly. That's the scenario that hurts, and it has nothing to do with whether Broadcom is well run.

Where I think this goes

I work in ranges rather than targets, so here's mine on a twelve-month view.

Bull case $520 to $580, which I'd put at about 45% likely. Base case $440 to $500, around 40%. Bear case $280 to $350, roughly 15%.

Weighted out that's an expected return somewhere near +35%, and it's the highest bull probability of anything I currently own. The whole position rests on AI capex continuing, which is also precisely what the bear case is about. I'm aware that's uncomfortable and I'd rather say it than dress it up.

What I'm doing about it: nothing. The contracts run to September 2027. I bought that much time specifically so a flat month wouldn't make me sell. If a 1% guidance miss breaks the thesis, there wasn't much of a thesis.

What I did do this week is start writing shorter-dated calls against some of my long-dated positions, so they earn something while I wait rather than only bleeding time value. Doesn't undo the 13%. Does mean the money is working.

Two things I'd like other views on.

Does anyone else here buy long-dated calls instead of shares? I find most people either don't know it's an option or have been burned once and never went back. I'd be curious which camp you're in and why.

And on Broadcom specifically — is 55% of revenue from five customers a risk you'd accept for that growth rate, or is it the thing that eventually matters? I've taken one side of that and I'm not certain about it.

Stop guessing. Start calculating.

Live to fight another day. 🤙