Oh, Fermi (FRMI).
How far you have fallen.
I first wrote about the data-center startup on October 1, right as it had its big, splashy stock market debut. Fermi ended up selling 32.5 million shares – more than it originally intended – due to high demand in this AI-crazy world we live in.
At $21 per share, which was at the higher end of its expected $18 to $22 range, it brought in gross proceeds of $83 million. That’s a pretty tidy sum for a company that was founded only seven months prior.
Moreover, as I mentioned in my article:
Fermi’s portfolio will consist of a single, massive 6,000-acre energy and data-center campus within 12 years. It’s currently being built within the Texas Tech University system outside Amarillo, Texas, on a ground lease.
When completely finished, the facility should be able to house 18 million square feet of AI-specific data centers, powered by nuclear energy, natural gas, wind, and solar power. In the shorter term, Fermi is aiming to have 1.1 gigawatts of power online by the end of next year.
Notice the phrases “will consist of,” “currently being built,” and “should be able to.” That last entire line about “shorter term” goals is worth highlighting, too, since it emphasizes how Fermi has no operational assets.
None.
Zip. Zilch. Zero.
And it won’t be for another year.
But, again, it’s a data-center company (or at least will be eventually). Plus, it’s co-founded by former Texas Governor and Energy Secretary Rick Perry. And it’s located near major natural gas pipelines; fiber data lines; and the U.S. Department of Energy’s Pantex plant, which is known as “the nation’s primary nuclear weapons center.”
Don’t get me wrong. Those are impressive associations and attributes.
It’s just that none of them could keep the stock from crashing 33.84% on Friday after one bad break.
An Utterly Unprofitable Data Center So Far
Here’s another problem with Fermi: It doesn’t actually have any customers.
It’s in talks with several hyperscalers. However, none of them have actually signed their names on the dotted line so far.
Worse yet, as we found out on Friday, one of them just terminated its $150 million Advance in Aid of Construction agreement.
Don’t ask me which one, because we don’t know. Since nothing was officially signed, nothing needs to be officially reported. All we know is that Fermi is still negotiating with the company about the larger potential lease.
Not surprisingly, that information didn’t keep investors from running for the hills. The stock sank $5.16 on the news. And since it has been falling almost every day since its first triumphant trading day…
That took it all the way down to $10.09 at Friday’s close. All told, the stock is down about 62% during its (admittedly brief) time as a public company.
Now, it probably didn’t help that the much larger, much more hyped-up AI play Oracle (ORCL) experienced its own negative news coverage at the exact same time. It had already fallen short of analysts’ lofty quarterly expectations on Wednesday. Then, two days later, Bloomberg reported that Oracle had delayed completion of some data centers for OpenAI by a full year.
It denied those allegations, saying that “all milestones remain on track.” But the damage was done, and not just to Oracle, which fell $8.88, or 4.47%, by the closing bell.
Alphabet (GOOG), Meta Platforms (META), and Amazon (AMZN), all major AI investors themselves, also fell notable amounts. The larger Nasdaq, meanwhile, tumbled 398.73 points, or 1.69%. And the S&P 500 lost 75.59 points, or 1.07%.
So, again, Fermi’s bad news came out at the exact wrong time.
Of course, there’s very little chance the aforementioned Big Tech players won’t bounce back again soon enough. Maybe even by the end of the year.
I even believe that Fermi will recover… eventually.
But it’s probably going to have to prove a thing or two before it does.
Consistency Is Key
Look, Fermi has a lot going for it, starting with the fact that it’s a data center (to be) provider.
I’ve detailed in this article how necessary these facilities already are and how much bigger they can still get. The ongoing AI revolution will cost trillion of dollars; and it cannot exist without data centers, which are the workhorses for every AI model.
That’s why I completely understand how excited investors got about Fermi. As I wrote on October 1, I was “intrigued” by the company.
However, I followed that up immediately with:
But it’s clearly in the very early stages of its existence, so I can’t recommend it today…
For the time being, we’ll just add it to our Wide Moat coverage spectrum. I always want to have at least several years’ worth of operational data to analyze, so let’s consider this the starting point of that initial investigation.
It should be interesting to see what happens from here!
And so it certainly has been.
Fermi’s rough patch is to be expected, though. New companies have a lot of kinks to work out. They’re still trying to figure out everything from their customers to their cash flow. So, starts and setbacks are to be expected.
What’s curious in Fermi’s case is that it’s going through these growing pains in full view of the public markets. Typically, a company like this would wait until it’s more established to IPO. And that means the stock investors may still have some waiting to do before consistent fundamentals develop (assuming they do).
By the time Fermi completes its campus a decade or so from now, it could very well be an absolute powerhouse with a market capitalization in the $70 billion-plus range. After all, Equinix (EQIX), the world’s largest data-center holder, carries that approximate market cap as I write.
But that’s a long road Fermi has to walk to get there. And there’s bound to be roadblocks and setbacks along the way.
That’s why we’ll continue to watch Fermi American. And if its underlying fundamentals ever warrant an investment – hopefully somewhere around its fourth or fifth year – I’ll let you know.
In the meantime, though, let’s leave the chaotic early stages to Fermi to figure out. We’ll find more peaceful paths to profits until it’s got something real to stand on.
Regards,
Brad Thomas
Editor, Wide Moat Daily
|