Avoid AI Data Newbie Fermi

AI mania is here, and readers need to sort through the froth.

We are at the stage when companies are trying to get a piece of the AI data center business, with almost no business behind the actual name.

Readers need to do analysis with a fine-tooth comb to sort out the winners from the losers.

Frenzies can be dangerous, and the allure of quick profits is often a mirage.

Look no further than the most recent AI data center that went public, which appears to be just smoke and mirrors.

Let’s talk about Fermi (FRMI).

Co-founded in January 2025 by former Texas Governor and U.S. Energy Secretary Rick Perry, the company raised $682.5 million by selling 32.5 million shares at $21 each, valuing it at a staggering $12.46 billion on debut—skyrocketing to $14.8 billion after shares surged 55% on opening day.

Is this the real deal?

In an era where AI’s voracious energy appetite could devour 21% of global electricity by 2030, Fermi positions itself as the ultimate “power behind the AI.”

But peel back the layers, and the hyper doesn’t match up.

At under a year old, with zero revenue, mounting losses, and execution hurdles that could sink lesser empires, Fermi embodies every classic red flag of a bubble-fueled IPO.

Here’s why you should steer clear, unless you’re a thrill-seeker with infinite patience and a tolerance for near-certain dilution.

The S-1 filing bluntly admits: “We are a development-stage company with no operating history or historical revenue, and we face execution risk across all major components of our business.”

At $13 billion, the market implies $1.2 million per megawatt (MW) for its aspirational 11 GW pipeline—rich even for proven players. Comparable AI data center operators like CoreWeave (which tripled post-IPO) or established REITs like Digital Realty trade at fractions of this premium because they have leases, cash flows, and tenants.

The revenue estimates are pitiful by industry standards.

A modest $1.5 billion in revenue from its first 1 GW by late 2026, and that has to be a huge letdown. 

This shovel-ready company is nothing to write home about.

They rushed to do their IPO, hoping that investors would bite.

They are far away from proving anything, and readers should take a wait-and-see attitude to this type of speculative company.

The suspect S-1 contents claiming that they are in the development stage has to force investors to just let this one find price discovery by itself.

I definitely won’t say this is a scam, but it appears as if this company is saying they are an AI data center company without much justification behind it.

It all smells like the run-up to the dot-com bust.

More and more, I hear anecdotally, companies are using the word AI to justify shrinking workforces just to get rid of labor expenses, while not integrating AI at all.

Don’t touch this one with a 10-foot pole. It screams high risk with low reward.