TLDR
- Burry calls AI’s language-first approach a “bad start” and a “known flaw”
- He introduced “Ballard’s Test” — true intelligence requires reasoning without language
- Burry argues the industry is stuck in a “parameter trap,” just scaling existing flawed models
- He sees a direct conflict between what Nvidia needs and what hyperscalers want
- Burry has placed new short bets against Nvidia, Tesla, and a semiconductor ETF
Michael Burry, the investor famous for predicting the 2008 housing crash, has taken aim at artificial intelligence — and the companies powering it.
In a post on his Substack page “Cassandra Unchained,” Burry laid out two separate but connected arguments: one about how AI is built, and one about how AI is funded.
AI Started on the Wrong Foot, Burry Says
Burry introduced a concept he calls “Ballard’s Test.” The idea is that genuine intelligence requires the ability to reason without using language at all.
He says the original goal of AI research was to build that kind of real reasoning first. But when that proved too hard, researchers switched to language-based models instead.
Burry calls this switch a “known flaw” and a “bad start.” In his view, the industry never fixed its foundation — it just moved on.
He says the result is a “parameter trap.” Instead of solving the core problem, companies are simply building bigger and bigger versions of the same flawed system.
He also pointed out that this approach depends on massive amounts of computing power — “zillions of power-hungry chips,” as he put it.
The Nvidia Problem
Burry then turned to the business side of AI, and he sees a conflict at the heart of the trade.
Nvidia needs AI chip demand to keep growing indefinitely. That’s what justifies its current revenue and the premium investors are paying for the stock.
Hyperscalers — companies like Meta, Amazon, and Microsoft — need the opposite. They want the heavy spending phase to end in three to four years so their costs come down.
“The hyperscalers are promising permanent demand growth and temporary spending over 3-4 years all in the same breath,” Burry wrote.
He says these two things cannot both be true at the same time.
Burry also flagged that free cash flow at major hyperscalers is already falling toward zero. Their reported profits look fine, but that’s partly because of long depreciation schedules masking real costs.
He says AI bulls imagine a “third door” — a future where demand stays high and spending drops, making everyone a winner.
Burry is direct: “There is no third door.”
He recently backed up his words with action, placing short bets against Nvidia, Tesla, and the iShares Semiconductor ETF.
Whether his timing is right or not, his argument raises a clear question — who actually wins in AI, and can both chip makers and hyperscalers come out ahead at the same time?
Burry says no.
Stop guessing and start investing with confidence. KnockoutStocks gives you the AI insights, market intelligence, and stock research you need to spot opportunities, cut through the noise, and make smarter investment decisions — all in one powerful platform.
Sign up today and get 50% OFF full access to our premium stock picks.
Simply use coupon code SPECIAL50 at checkout to claim your exclusive discount.
Get 3 Free Stock Ebooks
Discover top-performing stocks in AI, Crypto, and Technology with expert analysis.
- Top 10 AI Stocks - Leading AI companies
- Top 10 Crypto Stocks - Blockchain leaders
- Top 10 Tech Stocks - Tech giants
