Post

You've probably heard recently that NVIDIA is claiming it's not Enron - and it's right! But it's very strange that at the end of November it decided to tell short sellers, in detail, that it isn't Enron, WorldCom or Lucent, in a massive four-page note. www.wheresyoured.at/nvidia-isnt-...

In the analyst note, NVIDIA got defensive, claiming it was nothing like WorldCom - and again, it's right, WorldCom was a massive fraud that overstated earnings by saying its operating expenses were capital expenditures, spreading immediate costs over years. www.wheresyoured.at/nvidia-isnt-...

NVIDIA also said it wasn't Lucent Technologies, which is, again, *correct* but not a comparison you want to defend yourself over. Lucent loaned money to customers to buy its stuff, when NVIDIA...invests in companies and helps them find financing to buy GPUs. www.wheresyoured.at/nvidia-isnt-...

So, is NVIDIA doing Enron stuff? Not really. Enron was fraudulent, top to bottom, using dodgy SPVs to move around debt and marking unpaid, non-existent deals as revenue. I advise you read this so that you understand how messed up Enron was. www.wheresyoured.at/nvidia-isnt-...

NVIDIA is a weird company, with 88%+ of its revenue coming from making GPUs for AI. It holds a virtual monopoly over this kind of GPU due to its own coding language/library CUDA, which allows you to run software on graphics processors. www.wheresyoured.at/nvidia-isnt-...

You may be wondering why *so many companies* are buying GPUs, given the lack of returns or reliability or ability to do much with LLMs. The answer is simple: FOMO, and the fact that big tech has run out of hypergrowth ideas like the smartphone or cloud computing. www.wheresyoured.at/nvidia-isnt-...

This desperation for growth has them astray - Meta, Google, Amazon and Microsoft need $2tr in AI revenue by 2030 or they've wasted their capex, due to the massive cost of data centers and constant decay in value of GPUs, making profits impossible to achieve. www.wheresyoured.at/nvidia-isnt-...

Another problem: depreciation. NVIDIA is on a yearly upgrade cycle, meaning that the revenue potential of a GPU declines constantly, meaning that the revenue potential of these chips is under constant price pressure. They lose both value *and* potential over time. www.wheresyoured.at/nvidia-isnt-...

Big tech really has no way to make back the money it spent on GPUs. Data centers take years and billions of dollars to build, creating a giant hole of costs to fill before you even turn them on - and on a yearly cycle, every GPU loses value the second you buy it. www.wheresyoured.at/nvidia-isnt-...

A brewing scandal: NVIDIA claimed to have "shipped 6 million Blackwell GPUs" in the last year, an estimated 10-12GW of IT load (power used). The data centers are still being built and don't have the power, meaning millions of GPUs are likely sitting in warehouses. www.wheresyoured.at/nvidia-isnt-...

8/12/25, 5:42 pm13 reposts73 likes

dinglius-bonkler.bsky.social

It no longer makes sense to buy NVIDIA's AI GPUs. The space and power doesn't exist to turn them on, providing AI compute and LLM services is deeply unprofitable with no end in sight, and demand for these services simply doesn't exist at scale. It's time to stop. www.wheresyoured.at/nvidia-isnt-...

@dinglius-bonkler.bsky.social

I am deeply worried for the future. NVIDIA's revenue is a chain of pain - selling GPUs bought with debt that are rented to unprofitable LLM companies subsidized by VC who sell access to unprofitable AI companies subsidized by VC. The end will be gruesome. www.wheresyoured.at/nvidia-isnt-...

Dave Mueller

@dave-mueller.bsky.social

9 Dec 2025

This was great! I knew of Enron, of course, but didn't really know that many of the details. That was fascinating.

Piston Hondo my beloved

@pistonhondofan.bsky.social

True, the only good thing that will come out from this the companies colapsing and having to abandon their data centers. To bad it also comes with the collapse of the economy :(

Uncle B

@keyoke420.bsky.social

The end will be gruesome, but necessary. A.I. needs to be destroyed, outlawed and highly regulated like fentanyl.

ageofriesen.bsky.social

@ageofriesen.bsky.social

This is a great piece. Thanks for writing it and sharing it outside of the paywall.

Nolan T, Jones

@nolantjones.bsky.social

@eightxeroxedbutts.bsky.social

To ask the most cynical of survival questions: what economies will be least impacted by the impending disaster?

Stork

The Gallinari Configuration

VCs and AI companies are working diligently, feverishly even, to ensure that as few economies as possible are left unaffected. Look around: supply of AI is everywhere, demand is near-zero. "Nobody wants AI, which doesn't work."

@weigard.bsky.social

@psionicerudite.bsky.social

I joined Groupon right around the end of its massive hiring spree. At our orientation, an executive proudly told an auditorium of people that this is how they classified ad spending.

pghmike.bsky.social

@pghmike.bsky.social

9 Dec 2025

I can't help thinking I'm watching a star after reaching the end of hydrogen burning (VC funding). Perhaps it will find more fuel in helium, oxygen, etc (customer demand boosted by investments from Nvidia and other infrastructure players). But we all know how this ends.

Psionic Erudite

The laws of physics will be enforced

Look, it's like you don't even get this, so let me repeat it. Number. Go. Up. It will keep going up forever. Infinite money forever. Infinite growth forever. What part of this do you not understand?

@physicsanddata.bsky.social