Huangsgiving Gets Canceled After AMD Drops $8.2B on World Labs to Chase Jensen Into Physical AI

By Stocks News   |   1 month ago   |   Stock Market News
Huangsgiving Gets Canceled After AMD Drops $8.2B on World Labs to Chase Jensen Into Physical AI

Huangsgiving beef > turkey. 

While you lie awake at night pondering if you should throw more of your savings account into stonks, or earn a measly 5% on a 5-year T-Bill…

Jensen Huang’s once removed first cousin just threw down a tractor trailer’s load of $100 bills for an ex-Google employee’s startup.

Last night, AMD announced it is acquiring World Labs, the AI startup founded by legendary Stanford professor and former Google AI chief Fei-Fei Li, for $8.2 billion in stock.

“Somebody take Lisa Su’s Robinhood password away.” -every AMD shareholder

For perspective, World Labs was founded in 2024.

Meaning this baby went from “new startup” to an $8.2 billion exit in the amount of time it takes the average American to pay off a Ford F-150 Nissan Altima.

And apparently Lisa Su didn’t need much convincing.

AMD already invested in World Labs and began working with the company last year on training and inference optimization.

Now Lisa has decided sharing custody isn’t enough.

She wants the whole damn thing.

The deal will bring World Labs founder Fei-Fei Li directly into AMD as executive vice president and chief scientist, reporting to Su.

And if that name sounds familiar to anyone who spent their formative years doing something more productive than watching SpongeBob reruns…

Li is one of the godmothers of modern artificial intelligence.

She helped create ImageNet, the massive visual database that played a foundational role in the deep-learning revolution, later served as chief scientist of AI and machine learning at Google Cloud, and currently teaches computer science at Stanford.

So what exactly did $8.2 billion buy?

Something called a “world model.”

WTF in the ChatGPT is that?

Most of the AI you and I use today is really good at understanding and generating information.

Words. Pictures. Videos.

World models are attempting something significantly more ambitious: Teaching AI how reality works.

World Labs’ models are designed to understand 3D space, objects, movement and the relationships between them, allowing AI systems to create and interact with simulated environments.

Its first major model, Marble, can take a handful of images and turn them into explorable 3D worlds.

Which is admittedly pretty sick.

But the real money shot isn’t letting you turn your honeymoon photos into some neckbeard’s Ready Player One fantasy. 

The big kahuna here is robots, autonomous vehicles, factories and warehouses.

You know, the kind of disruption that actually affects the physical AI economy.

For instance, supposedly, world models could allow robots and other AI systems to train inside realistic simulations before getting released into the physical world.

Translation? Instead of teaching Optimus how stairs work by watching a $30,000 robot eat sh*t down a flight of them…

You let him break his imaginary neck 40 million times inside a computer first.

Progress.

And THAT is where this goes from “Lisa, what the f*ck?” to “Ohhhhh.” 

World Labs can show AMD where AI hardware needs to go years before the rest of the market gets there.

As World Labs develops increasingly complicated models for robotics, simulation and physical AI, its researchers will inevitably discover where today’s hardware is flawed.

What does that mean? I’m glad you asked.

More memory. More compute. Faster inference. Different architectures. Whatever other black magic Jensen currently keeps locked underneath Nvidia headquarters.

AMD gets to see those problems firsthand. Then it gets to build the chips designed to solve them.

Ohhhhhhh.

Now the $8.2 billion tractor trailer starts making sense.

Instead of waiting for tomorrow’s AI companies to tell AMD what hardware they need…

Lisa Su just bought tomorrow’s customer.

And more importantly, she bought the research lab that can help tell her what tomorrow’s customer is going to need.

That matters when your biggest competitor happens to be Nvidia.

Speaking of Cousin Jensen, the man already has AI compute by the balls and has now turned his attention toward giving the machines arms and legs.

Nvidia already has its Cosmos world models cooking up fake roads, nightmare traffic scenarios and other assorted ways to teach autonomous cars not to kill us. 

Add in its robotics platforms, simulation tools and approximately 47 other products Jensen keeps pulling out of his leather jacket like Mary Poppins.

I think you get the picture.

AMD has some catching up to do, and Lisa Su apparently decided patience is for companies without $8.2 billion lying around. 

Next family get together’s about to be awkward as sh*t.

At the time of publishing this article, Stocks.News holds positions in Google, Robinhood, and Ford as mentioned in the article.

Did you find this insightful?

Disclaimer: Information provided is for informational purposes only, not investment advice. We do not recommend buying or selling stocks. Stock price discussions are based on publicly available data. Readers should conduct their own research or consult a financial advisor before investing. Owners of this site have current positions in stocks mentioned throughout the site, Please Read Full Disclaimer for details Here https://app.stocks.news/page/disclaimer