# AI chatbots bleed, the picks and shovels get paid

By @FrankCurzio on X · https://f.xyz/t/43223a02-3df1-405d-9f48-5513a6bb938f

> Whoever builds the clouds, chips, data centers and power for AI makes money, even if chatbots like Grok lose money.

**Result:** If held over the last 3 months (Aug 12, 2026 – Nov 11, 2026): −20.8%. Over the same days: Chip stocks −15.8%, S&P 500 −2.4%.

## What it holds

- **Big clouds** (30%, −5.3% over 3 months): AMZN (Amazon) −1.5%, MSFT (Microsoft) +8.2%, GOOGL (Alphabet (Google)) −16.0%, ORCL (Oracle) −18.8%
- **Chip makers** (25%, −16.0% over 3 months): NVDA (Nvidia) −18.3%, TSM (Taiwan Semiconductor) −7.8%, AVGO (Broadcom) −24.1%, MU (Micron) −16.1%
- **Data center builders** (20%, −42.8% over 3 months): CRWV (CoreWeave) −53.5%, NBIS (Nebius) −33.0%, IREN −49.0%, DELL (Dell) −22.5%
- **Power and cooling** (15%, −14.8% over 3 months): GEV (GE Vernova) −21.4%, BE (Bloom Energy) −10.9%, VST (Vistra) −9.7%
- **Data center wiring** (10%, −31.6% over 3 months): MRVL (Marvell) −29.2%, CIEN (Ciena) −31.6%, COHR (Coherent) −35.1%, LITE (Lumentum) −30.6%

## The plan behind it

Every dollar a chatbot burns goes to someone else: cloud rent, chips, power and buildings. So this portfolio owns the ones getting paid, not the chatbots. The big clouds are the steady winners, because AI runs on their servers. The chip makers and the data center builders are the higher-risk, higher-reward part. This includes CoreWeave, which has fallen hard even though orders are still coming in. I left out shorts because the take is a long bet, and the weak borrowers he warns about are not on the list.

> Elon just admitted Grok costs more to run than it brings in.
> 
> Which raises the question...
> 
> If AI loses this much money, why is everyone still pouring trillions into it?
> 
> After my post on Saturday, someone asked me exactly that.
> 
> Fair question, but there's a reason every company is still doing it. And let me tell you...
> 
> Losing money early isn't new.
> 
> Amazon didn't post a full-year profit until 2003, nine years after it started. Everyone called it a joke. I remember being on Wall Street watching it happen.
> 
> And it's a similar story now with AI.
> 
> The giants spending the most on AI aren't doing it for the chatbot. AI runs on their clouds, and cloud is one of the most profitable businesses on earth. Amazon's runs at margins near 40%. Every company that wants AI has to rent it from them.
> 
> More importantly, nobody wants to be the one who sat this out either.
> 
> Whoever owns the AI infrastructure for the next 20 years wins, so they'd rather overspend now than lose the whole damn thing.
> 
> And it gets cheaper. Every technology starts out expensive and clunky. The cost of running these models keeps falling, and whoever builds the infrastructure now makes the money when it does.
> 
> Does that mean everyone wins? Hell no.
> 
> The guys who borrowed billions with no real customers are going to get crushed, and some won't make it. And the companies that fired half their staff thinking AI could run itself are already learning that the hard way.
> 
> Look at Elon's own numbers. Grok loses money, but every dollar it burns goes to someone: the chips, the power, the data centers. And that's the best part.
> 
> Some will think that's fucking absurd, because a lot of these stocks have been annihilated in recent weeks.
> 
> CoreWeave is down over 70% from its high. Vertiv is down about half... BUT the money is still flowing to them.
> 
> Vertiv’s orders more than doubled last quarter, and CoreWeave has over $100B of signed business. What got crushed was the price, because the market panicked about who’s going to
> — @FrankCurzio on X: https://x.com/FrankCurzio/status/2120124957103681543

**What would prove it wrong:** Big cloud companies start cutting their AI spending plans, and orders at the chip and data center firms fall.

## FAQ

### How do I bet that Whoever builds the clouds, chips, data centers and power for AI makes money, even if chatbots like Grok lose money?

This tested idea holds AMZN (Amazon), MSFT (Microsoft), GOOGL (Alphabet (Google)), ORCL (Oracle), NVDA (Nvidia), TSM (Taiwan Semiconductor), AVGO (Broadcom), MU (Micron), CRWV (CoreWeave), NBIS (Nebius), IREN, DELL (Dell), GEV (GE Vernova), BE (Bloom Energy), VST (Vistra), MRVL (Marvell), CIEN (Ciena), COHR (Coherent) and LITE (Lumentum). Make a copy to run it on practice money first, at real prices.

### What would prove it wrong?

Big cloud companies start cutting their AI spending plans, and orders at the chip and data center firms fall.

### Can I make one from my own idea?

Yes. Write what you believe in plain words at https://f.xyz/create. fxyz asks a couple of questions, builds the portfolio and shows how it would have done over the last 3 months.

### Can I turn a post on X into a theory?

Yes. Reply to the post and tag @fxyzai. It tests the view in the post and replies with the 3-month result and a link.

### What can it trade?

Stocks (Tesla, Apple, Nvidia), crypto (Bitcoin, Ethereum, Solana), pre-IPO companies (OpenAI, Anthropic), indices (Nasdaq 100, S&P 500), commodities (gold, silver, oil, copper) and currencies (EUR/USD, USD/JPY). One theory can mix them, and bet for or against any of them.

### How does it react to news?

On each check it reads the last two days of news on its biggest holdings and says whether the idea still holds. It can make one small change at a time: trim or drop one holding by up to 5 points. On autopilot the change waits 10 minutes so the owner can skip it; otherwise it asks first.

### How do I run it with real money?

For real money, tap Make it mine, open your copy and choose Use real money, with an exchange connected: Hyperliquid, Lighter, Lighter (Robinhood), Backpack, Pacifica or Binance. The money stays in your own account there. fxyz can place trades, but it can never withdraw.

### Can the AI get it wrong?

Yes. fxyz's AI turns the idea into holdings, and it can get that wrong: add a coin or stock that doesn't fit, or miss one that does. Check the holdings before you put money in. In your own copy, ask its chat to add or drop any of them.

### What do the 3-month numbers mean?

What $100 would have done holding this mix over the last 3 months, at real prices. It is a what-if, not a record: past or simulated results are no promise of future ones, and nothing here is investment advice.

Test your own idea: https://f.xyz/test
