---
title: 'The AI Trade is Cracking.'
source: 'https://youtube.com/watch?v=nRfXu1341Yo'
video_id: 'nRfXu1341Yo'
date: 2026-11-16
duration_sec: 469
channel: 'Rule #1 Investing'
---

# The AI Trade is Cracking.

> Source: [The AI Trade is Cracking.](https://youtube.com/watch?v=nRfXu1341Yo)

## Summary

The video analyzes the recent decline in AI and semiconductor stocks, arguing that the AI trade was priced for perfection and is now cracking. It contrasts this with the historical example of Cisco and Ross Stores to illustrate the importance of buying wonderful businesses at a reasonable price.

### Key Points

- **AI Trade Decline** [00:02] — Memory stocks have lost significant value since early October, with Intel down 66%, Western Digital down 66%, SanDisk down 66%, ARM down 63%, Seagate and Marvell down 59%, Applied Materials and Lam Research down 57%, and NVIDIA down 25% from its high. The semiconductor index is down 40%.
- **Priced for Perfection** [00:46] — The speaker previously stated that AI chip prices only made sense if everything went perfectly, which rarely happens in investing. The decline was triggered by Intel warning of a PC slowdown extending into 2027, causing stocks to fall over 9% and then continue dropping.
- **Michael Burry's Depreciation Warning** [02:09] — Michael Burry highlights that Big Tech spreads the cost of AI chips over their useful life (depreciation), but hyperscalers have been stretching that life even as chips become obsolete faster. This understates expenses, inflating profits by roughly $176 billion from 2026 to 2028, which Burry calls 'edging up on investor fraud.'
- **Venture Capital Concentration** [03:04] — Burry notes that about 80% of US venture capital this year went into AI-linked companies, compared to only 40% into internet companies at the peak of the dot-com bubble. This suggests the AI bubble could burst sooner than expected.
- **Self-Reinforcing Decline** [03:18] — The decline feeds on itself: stocks fall, boards get nervous, companies cut data center plans, leading to fewer chip orders, earnings misses, and further stock declines. This mirrors Cisco, Lucent, and Nortel in 2000-2001, where customers stopped buying all at once.
- **Historical Parallel: 1999-2000** [04:03] — In 1999, the Nasdaq rose 86%, but Warren Buffett's Berkshire fell 20%, leading to criticism. However, over the next 5 years, the best-performing stocks had nothing to do with the internet. The internet did change everything, but the winners (Google, Facebook) came later, while Yahoo, Netscape, and AOL were the leaders then.
- **Ross Stores vs. Cisco** [05:12] — At the end of 1999, Ross Stores was a boring off-price retailer, while Cisco was the hottest tech stock. From 1997 to the 2000 peak, Cisco rose 700%, but then took 25 years to get back to that high. In contrast, $10,000 in Ross Stores on January 1, 2000 would be worth about $1.39 million today, while the same in Cisco would be about $28,000.
- **Investment Rule One** [06:27] — The speaker's rule is not to chase hype but to buy wonderful businesses you understand with a durable moat and great management at a price that gives a margin of safety, then have patience. Boring businesses are currently ignored because they don't mention AI on earnings calls, creating bargains.

### Conclusion

The AI trade is cracking due to overvaluation and accounting concerns, but the key lesson is to focus on buying wonderful businesses at reasonable prices, as exemplified by Ross Stores versus Cisco. The speaker invites viewers to identify the 'Ross Stores of this era' and learn valuation skills.

## Transcript

memory stocks have lost since early October by Yahoo Finance's account. And this summer, I sat right here and I told you the AI trade has gotten way ahead of record highs, I'm sure a lot of you
thought, well, Phil's an old value guy, he just doesn't get it. So let me show you what happened. Okay, Intel [clears throat] down 66% from its high, Western Digital down 66%, SanDisk down 66%, ARM down 63%, Seagate and Marvell
down 59%, Applied Materials, Lam Research 57%, NVIDIA. Best of the bunch is still 25% off the high it just set last month. All right, the semiconductor index down 40%. Now, I want to be honest with you guys, I didn't know that this
would happen specifically in October. My crystal ball isn't quite that good. What I said was that these prices only made sense if everything went perfectly, and in investing, perfect almost never shows up. Now stick with me because in a few
minutes, I'm going to show you a boring retailer that turned $10,000 into more than a million dollars while the hottest tech stock in America went nowhere for 25 years. That story is the most important lesson that I can teach
you right now. Here's what I said. These companies, specifically the chip equipment makers, were priced for perfection. And when I say that, what I mean is that in order for that stock price to be a rational price, their
growth rates can't slow down, their PE ratios have to stay at the moon. And for that to happen, every single thing has to be perfection for years. Every data center has to get built, every order has to show up on time. There's no
depreciation on old chips, nobody can slow their spending, not for a quarter, not ever. But oh oh, SOX, the semiconductor ETF, started slipping and slipping. And then on October 30th, Intel warned that a PC slowdown would
stretch out into 2027. It's not a disaster, you guys, it's just business will be a softer. That's all. And immediately stocks fell more than 9% and then in the next few days it kept falling and now it's down 23% in a week.
A little less than perfect. That's all it took. Now Michael Burry, the guy from the Big Short who called the 2008 housing crisis, has been hammering one point on his substack and every investor needs to understand it. When Big Tech
buys AI chips, it doesn't expense the cost all at once. It spreads it over the equipment's useful life. That's depreciation. Burry said the hyperscalers have been stretching that useful life even as the
chips go obsolete faster than ever. So the simple version of this is you buy a chip that's really useful for 3 years, you spread the cost over 6 and you only book half the expense each year. So profits that look way bigger than they
ah, they're a little bit of fiction. By Burry's math, that's roughly $176 billion in understated appreciation from 20 6 to 28. That's not perfection. That's edging up on investor fraud. Then in late September he said this bubble
could burst sooner than even he expected. He pointed out that about 80% of US venture capital this year went into AI linked companies. At the peak of the dot com bubble they only had about 40% in internet.
Burry's been early before, okay, but early and wrong are not the same thing. So here's what worries me the most. This kind of decline just feeds on itself. Stocks fall, boards get nervous. A company that planned three data centers,
they cut it to two and that cut flows straight up the chain. You get fewer chip orders, fewer equipment orders, you get an earnings miss and stocks fall get an earnings miss and stocks fall again and down begets down. And that's
Cisco, Lucent and Nortel in 2000 and 2001. The internet didn't go away. Their customers just stopped buying all at once. And I can't tell you how far this one goes, but history rhymes. And right now it's rhyming loud. In
now it's rhyming loud. In Nasdaq went up about 86% Warren Buffett, the world's greatest investor ever lived, watched Berkshire fall about 20%. Barron's put him on the cover. What's wrong, Warren? People said he'd lost it.
Value investors like Seth Klarman, they got called dinosaurs. Then sure enough, the script flipped. The market crashed. And amazingly, over the next 5 years, a lot of the best performing stocks had nothing to do with the internet. And
here's the thing, you guys. The internet did change everything. The bulls were right about that. They were just wrong about who would win. In 1997, the internet leaders were Yahoo, Netscape, and AOL. Google didn't even
exist until 1998. Facebook, not until 2004. Amazon was a money-losing online bookstore, and the real winners came mostly out of nowhere. AI is going to
change the world. I believe that, but can I say for sure 10 years out who the winners will be? No way. No way. So back in 1999, what did Buffett and Klarman buy? They bought wonderful businesses. They bought
businesses they understood. Let me show you what that looks like. At the end of 1999, Ross Stores were a boring off-price retailer. Nobody cared. In fact, between '97 and '99, Ross stock went up and then down 50% three
different times. Imagine holding that while your neighbor brags about his Cisco because Cisco was the hottest company in America, the Nvidia of its day. Picks and shovels for the internet. From 1997 to the 2000 peak, it went up
700%. Then it took 25 years just to get back Then it took 25 years just to get back to that high. A quarter century of dead money. And Ross? Well, you'd be up about 140 times your money today. Not 140% No,
140 times your money today. Not 140% No, 140 times. If you put 10,000 into each stock on January 1st, 2000 and sold today, your $10,000 of Ross stock would be worth about 1 million 390,000 including dividends, and Cisco about
28,000. Cisco's a great company. That was never the problem. The problem was the price. Back then you were paying upfront for 20 years of amazing projected growth and when reality set in, even a great company becomes dead
money. [music] So here's what we are doing. We're not chasing hype. Dead money is not rule one. Rule one is buying wonderful businesses you understand with a durable moat and great management at a price that gives you a
real margin of safety. Then having the patience to let the market catch up. [music] And right now, just like 1999, wonderful boring businesses are being ignored because they don't say AI on their earnings call. And that's fine
bargains are. So will we look dumb for a while? Probably. Buffett did. Our goal is to look a little dumb today and really, really smart 5 to 10 years from now. So no
guarantees, but that's the bet I'm making with my money. So here's my question for you guys. What are the Ross stores of this era? What is the boring wonderful business that everybody's ignoring? Please drop it in the comments
cuz I read them. And if you want to learn how to value these businesses yourself, you guys, what a company's worth, what price gives you a margin of safety, join me at the world's number one investing workshop. I'll be there
because I teach it. So link below and we'll see you at the next one.
