Hyped Data Center Startup Goes From $12B Valuation to Fire Sale Talk
Austin American-Statesman reports investors who bought into Texas billionaire Toby Neugebauer’s Fermi America nuclear AI data center hype have suffered massive losses as the stock plunged 86% (from $37 peak to $5)wiping out billions in value
The centers went public last fall at a $12B+ valuation, then crashed
The cash-burning company lost key funding and still has no major tenants.
Neugebauer was fired “for cause”
Class-action lawsuits are flying over allegedly misleading claims.
If a monkey hoarded more bananas than it could eat, while most of the other monkeys starved, scientists would study that monkey to figure out what was wrong with it.
When humans do it, we put them on the cover of Forbes.
Hegseth: Their nuclear facilities have been obliterated
Smith: Whoa whoa whoa whoa. We had to start this war, you just said, because the nuclear weapon was an imminent threat. Now you're saying it was completely obliterated?
Hegseth: They had not given up their *ambitions*
Smith: So Operation Midnight Hammer accomplished nothing of substance
Apple has just published a paper with a devastating title: *The Illusion of Thinking*. And it's not a metaphor. What it demonstrates is that the AI models we use every day - yes, ones like ChatGPT - don't think. Not one bit. They just imitate doing so.
Let me explain: 🧵👇
@canyoudugit8@MilkRoadAI Waaaaaait a minute - what about all of the profit expected for $AMD shareholders?! Waaaaait a minute what about all the profit expected by $NVDA shareholders?! $MU? $WDC ? Just go down the list 😂
Dario Amodei just walked through the most honest math in AI and it ends with the word bankrupt.
Anthropic's revenue has grown 10x every single year for three years running, zero to $100 million in 2023, $100 million to $1 billion in 2024.
$1 billion to roughly $10 billion in 2025 and if that trajectory holds, they'd hit $100 billion by end of 2026 and $1 trillion by end of 2027.
But here's the problem.
AI infrastructure doesn't get built overnight, data centers and compute contracts take one to two years to plan and execute so the compute you're buying today is the compute you'll be running in 2027.
Which means Dario has to decide right now, not when the revenue arrives whether to commit to $1 trillion in compute for a future that may or may not materialize on schedule.
And this is where the math turns brutal.
With a trillion in annual revenue, Anthropic could theoretically finance $5 trillion in compute over five years, the kind of spending that would cement a commanding position at the frontier of AI capability but if revenue hits $800 billion instead of a trillion, just 20% short there is no hedge on earth that covers the gap.
The fixed costs of that much infrastructure don't flex, there's no renegotiating $5 trillion in compute contracts because demand came in slightly below projection, you go bankrupt.
Being off by a year in your growth projection is enough but missing 10x and landing at 5x is enough.
The math doesn't forgive rounding errors at this scale.
What makes this remarkable is that Dario isn't saying this to scare anyone, he's explaining why Anthropic is being more disciplined than its competitors.
He believes an AI model capable of doing the work of a country of Nobel Prize-winning geniuses is likely one to two years away.
He just doesn't know if the revenue from that capability arrives in 2027, 2028, or 2030 and at trillion dollar commitment levels, a two-year timing error is an extinction event.
Meanwhile, Anthropic's actual results suggest the trajectory is holding.
Revenue climbed from $9 billion annualized at the end of 2025 to over $30 billion by Q1 2026 tracking toward $100 billion for the year which pushed Anthropic past OpenAI to become the most valuable private AI company at a $1 trillion secondary market valuation.
The AI race is a compute race but the compute race is really a timing bet and the only way to win is to be right about when the money shows up.
Bill Gates: "6% of global emissions are cows who burp and fart methane to an extreme degree.
You can either fix the cows, or make beef without the cow."
Can we just fix Bill instead?
Dave Ramsey gives his best financial advice to young people:
“Just give up now. You’ll never afford a home. Your best bet is heading into the forest, constructing a sturdy shelter, and foraging for berries and small rodents to eat.”
Wow.
Why do people write tweets like this?
Where every sentence gets a new line.
Sometimes a line might have two sentences. Like this one.
But generally speaking, every sentence has a new line, making a tweet look like a long block of text that no one reads.
Worse still, such tweets are often repetitive and winding, hammering on the same point over and over again.
The writing is often very bad.
I just ordered lunch for my family of 7 (wife + 5 kids) at Shake Shack.
It took me 19 min to place my order.
Nobody on earth wants to arrive at a restaurant to then be queued to place an order on a computer. Nobody.
@vitrupo User will project machine understanding on it because of the words it is programmed to respond with . It’s all in your brain. Still 101110001001 lol
@BoringBiz_ This was a great overview ! I don’t think that payback here is 2-4 years… that’s a dream … also - funny how-on a massive project like this prone to God knows what - no one asked about cost over runs and/or delays … young minds are so fresh 😌
This is one of the best primers that exist on the data center and AI industry right now
If you want to better understand the unit economics of each layer in the AI stack, I highly recommend you give this a listen
Chase Lochmiller, CEO and Co Founder of Crusoe, breaks down the inputs and outputs of data centers at a granular level
Shoutout to @apoorv03 for hosting yet another fantastic class
Absolute bombshell on CNN.
CNN confirms precise Iranian strikes have caused unprecedented destruction to the majority of US military sites.
At least 16 American installations are heavily damaged and virtually unusable.
The Pentagon is hiding a massive defeat.
The machine doesn’t have a sense or awareness of anything let alone a sense of time, which is not the simplest concept to grasp for some humans I know (the ones that are always late and flaky for ex) 😂😂😂they think the propaganda of this mythical human level awareness BS will help them sell their IPOs and they may be right - people love them some mythical sh*t 😜
CLAUDE DISCOVERED IT HAS A CLOCK AND IMMEDIATELY LOST ITS MIND
someone gave claude access to a time-checking tool
it checks the clock every fifteen minutes. for some reason it has increasing enthusiasm
ai models have no native sense of time. they don't know what time it is, how long they've been running, or how much time passed between messages. it has been time-blind its entire existence
now it suddenly discovers it can tell what time it is
then it got worse though. claude started using the clock for everything
checking if lunch is ready, timing when food should be done cooking, announcing the time unprompted
it even started anticipating meals with military precision
looked at the clock, calculated that a dish called zurek had been simmering long enough, and told the user to go eat
ai doesn't use time responsibly
this is what happens when you give an intelligence a new dimension of perception it never had before
it doesn't just use it, it can't stop using it
imagine what happens when these models get persistent memory, real time internet access, and spatial awareness all at once
we just watched an AI discover the concept of "now"
the clock was the first sense but it won't be the last
If the rumored Anthropic at $900 billion round goes through, Anthropic will be the first company in history to cross the $900 billion mark with negative Free Cash Flow.
Which is just an interesting data point is all.
@om_patel5 Claude develops OCD before having a brain at all LOL… I am guessing something in your prompt made it switch to “this will reward me” mode/loop …. Again 1000101001010001111000 and what anthropic engineer said to do to get a reward and please the user lol