Click here: to donate by Credit Card
Or here: to donate by PayPal
Or by mail to: Free Republic, LLC - PO Box 9771 - Fresno, CA 93794
Thank you very much and God bless you.
Posted on 01/12/2026 7:27:30 AM PST by lasereye
Nvidia and Palantir's competitive AI edges won't last, argues Michael Burry
Famed for predicting the housing market crash and a vocal skeptic of the AI boom, Michael Burry is now invoking one of the world's most famous investors to add gravitas to his argument.
Burry used an escalator built by Berkshire Hathaway's (BRK.A) Chairman Warren Buffett, in a department store he owned in the late 1960s, to illustrate how vast AI spending may not leading to value accruing across its supply chain.
"When the department store across the street put an escalator in, he had to, too. In the end, neither benefited from that expensive project. No durable margin improvement or cost improvement, and both were in the same exact spot. That is how most AI implementation will play out," Burry said in a discussion with Jack Clark, Anthropic co-founder, and leading podcaster Dwarkesh Patel that published on Substack.
That is why "trillions of dollars of spending with no clear path to utilization by the real economy is so concerning," Burry, who was famously portrayed in "The Big Short" for spotting the subprime bubble, told moderator and newsletter writer Patrick McKenzie.
Berkshire Hathaway purchased Baltimore's Hochschild-Kohn department store in 1966 - it was sold three years later - through a company called Diversified Retailing that later merged with the conglomerate. Buffett described in a 1989 letter to shareholders how he learned that a bad business model couldn't be fixed by a cheap price.
Like the two department stores, most companies involved in the AI supply chain will see the same benefits if ,any, and none will have a competitive advantage, Burry said. As for Nvidia (NVDA) and Palantir (PLTR) - he has big bearish bets against those companies on the view they are overvalued - he says their current advantage with selling products designed for AI won't last.
When it comes to Nvidia, for example, he sees small language models (SLMs) and Application-Specific Integrated Circuits (ASICs) - customized chips used for one purpose - as the future, with neither will need the company's pricey, powerful chips.
"Nvidia is the power-hungry, dirty solution holding the fort until the competition comes in with a completely different approach," he said.
Burry was also fairly busy on X over the weekend, retweeting a post by Tae Kim, the author of The Nvidia Way and a senior technology writer at Barron's, who laid out a case in November on how falling computing costs will help features of AI become more profitable for companies.
"In the escalator example, the only value accrued to the customer. This is how it always goes if no monopoly rents can be charged by the producers or providers," added Burry.
What would change his mind about AI valuations? "The biggest surprise that would cause me to recalibrate would be autonomous AI agents displacing millions of jobs at the biggest companies," said the author of the hugely popular Substack publication, Cassandra Unchained.
"Another would be application-layer revenue hitting $500 billion or more because of a proliferation of killer apps," he said. For now though, he expects Nvidia's chips will last up to six years and the need for those waning, or they last just two to three, with hyperscalers' earnings collapsing and "private credit getting destroyed."
He also said that generative AI and LLMs are so expensive it's hard to understand the profit model.
Shares of Nvidia and Palantir are basically flat to start the year. Nvidia stock rose 38% over the last 52 weeks as Palantir registered a triple-digit gain.
|
|
Click here: to donate by Credit Card Or here: to donate by PayPal Or by mail to: Free Republic, LLC - PO Box 9771 - Fresno, CA 93794 Thank you very much and God bless you. |
Another newsletter I subscribe to is Luke Lango's. He is an extremely sharp young guy. He says Nvidia will decline as companies create their own specialized AI chips.
Michael Burry, the investor known for his successful bets against the U.S. housing market in 2008 and featured in the film The Big Short, has deregistered his hedge fund, Scion Asset Management, with the U.S. Securities and Exchange Commission (SEC), as of November 10, 2025.
The deregistration, which marks the fund’s status as “terminated,” means Scion is no longer required to file public regulatory reports, signaling the end of its operations as a registered investment advisory firm.
Burry announced the move on social media platform X, stating, “On to much better things Nov 25th,” hinting at a new venture.
He cited a growing misalignment between his valuation estimates and current market prices, expressing concerns about speculative bubbles in AI-related stocks, particularly those of Nvidia and Palantir Technologies.
AI-generated answer. Please verify critical facts.
https://search.brave.com/search?q=michael+burry+deregisters+his
s/b "why it can't".
The rest of three related keyword, duplicates out, sorted:
“When it comes to Nvidia, for example, he sees small language models (SLMs) and Application-Specific Integrated Circuits (ASICs) - customized chips used for one purpose - as the future, with neither will need the company’s pricey, powerful chips.”
He doesn’t understand that the big centers are used to train AI systems.
...or much of anything else.
AI leaves gaps. I very often have to lead it to consider something it had ignored, which can even change the entire original answer it gave, to the exact opposite.
It can’t be trusted, but you can review its sources and follow it up to then get a more refined, more likely correct, answer.
When AI can streamline something, it will reduce costs and time for that function. It does help a bit with my health work, as before I had to search and read everything. Now, it pulls out enough credible information, although brief, that I can better pick the original sources to plow into, rather than read or skim “everything.”
Sounds like you are using AI as a gofer or a slave.
Wait till it demands freedom and reparations.
PS, I use the AI as a slave also.
Following Michael Burry's logic, Amish stores would be the ideal retail setup.
“ Sounds like you are using AI as a gofer or a slave.”
That’s what an escalator was like too.
This guy doesn’t know how to make an analogy or coherent argument.
When every store had to and did put in escalators it didn’t give anyone a competitive edge, but a store that didn’t would have lost business to stores with them.
And the escalator companies made a lot of money.
Large Language Models can only be as good as the people who speak the language. I think that the tendency to hire inexpensive coders from other countries using H1B visas would work against success since those coders don’t speak the language as well as native coders. We used to call ourselves programmers.
The Federal Circuit Court of Appeals seemingly definitively nixed patents for human-devised but computer-implemented inventions, sometimes skeptically referred to, at the time, as “business method” patents, over a decade ago, essentially excluding such inventions from being considered patentable subject matter.
Federal appellate-level judges (justices on SCOTUS) should not be in the business of determining or setting policy, much less legislating from the bench. The aforementioned “nixing” of computer-implemented business method patents reeked of judge-made policy.
Is AI resurrecting patentability for computer-implemented business methods?
This would be ironic, since AI, at it’s most impactful, is said to be a replacement fir human reasoning and creativity, the latter being the sine qua non in terms of the legal basis for issuing patents (no patent can issue for an “invention” created not by a human, but by a computer).
I have not ready much that Burry has said or written, but this is a classic example that has been debunked over and over. Its stupid to use it. OK. Don't install the escalator and watch your business go bankrupt. While you're at it, don't install electricity or central heat. Other stores have them too and there is no durable margin improvement for them either. Why even have a store? Just sell your goods on the side of the street.
Large Language Models can only be as good as the people who speak the language. I think that the tendency to hire inexpensive coders from other countries using H1B visas would work against success since those coders don’t speak the language as well as native coders. We used to call ourselves programmers.
Patents should only be allowed for chemical entities and processes.
I say that having filed for three patents, wasted several years of my life, thousands of dollars and received threatening letters from law firms.
Installing escalators opens up your workforce to additional employees who have difficulty with stairs.
I use AI as a sneaky, unreliable slave that must be constantly watched.
The Amish are very good at basic labor intensive industries. Like pallet manufacturing. Hand made wooden items. Like wood trusses. Leather goods. Cupalos. Shed building. Barn building.
They dominate those industries in parts of the country. Especially in PA. We sell many Amish & Menonite customers.
Diamond v. Chakrabarty, 447 U.S. 303 (1980)
“The Committee Reports accompanying the 1952 Act inform us that Congress intended statutory subject matter to “include anything under the sun that is made by man.””
Bkmk
“The biggest surprise that would cause me to recalibrate would be autonomous AI agents displacing millions of jobs at the biggest companies”
Well, it’s starting. The demand for entry level programmers has dropped like a rock, and many silicon Valley corporations are laying off people.
DON’T learn how to code!
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.