AI: Now With Less Circle And More Jerk

Did the AI capex bubble just burst before the weekend - while Iran and private credit already have the market volatile?

As if markets didn’t already have enough to worry about heading into the weekend — an escalating conflict involving Iran, growing stress in private credit, and the ongoing annoyance of positive real interest rates — one of the “AI will solve everything, just add capex” deals that everyone in markets have been quietly laughing about while investing in for the past year has officially started to unravel.

According to Bloomberg, Oracle Corporation and OpenAI have scrapped plans to expand a flagship artificial intelligence data center campus in Abilene, Texas after negotiations dragged on over financing and, more awkwardly, OpenAI’s “changing needs.”

Changing needs, of course, being corporate-speak for: the numbers probably stopped making sense once someone sat down with a spreadsheet and noticed that data center financing deals are starting to arrive dead on the operating table.

The project in question sits in Abilene and is being developed by Crusoe Energy Systems as part of the highly publicized Stargate initiative — one of the many AI infrastructure megaprojects that have been breathlessly announced over the past year with the implicit assumption that demand for compute will grow forever, financing will always be available, and electricity will somehow materialize in gigawatt quantities on command.

The site itself is enormous: roughly 1,000 acres of land designed to host hyperscale data center clusters that would consume multiple gigawatts of power. Portions of the facility are already operating, and construction continues. But the big expansion — the one that was supposed to anchor the next phase of the project — suddenly no longer has a tenant.

Which is not exactly the kind of development you want in the middle of what’s supposedly the most unstoppable technology boom since the internet.

The timing of the headline didn’t exactly help market nerves either. The news crossed the tape around 3:00 PM EST, just as traders were already digesting a fairly ugly macro backdrop, and equities promptly faded into the close.

The late-day selling wasn’t catastrophic, but it was noticeable — exactly the kind of “huh, that’s interesting” price action that tends to get revisited when markets reopen after a weekend full of geopolitical headlines. In other words, if there’s follow-through pain on Monday, this little AI infrastructure hiccup will likely be part of the blame cocktail. The other part of the cocktail will likely be a result of the ugly headline about gating redemptions from a $26 billion fund that crossed the wire early this morning.

Private Credit Just Turned Into Public Panic

Into the suddenly vacant Abilene expansion steps Meta Platforms, which is now reportedly considering leasing the space. And because this is the AI industrial complex we’re talking about, the story quickly gets more entertaining.

Nvidia — the undisputed king of the AI gold rush — apparently helped facilitate the discussions between Crusoe and Meta after the Oracle/OpenAI expansion fell apart. The reason is simple: if the expansion stalled, there was a non-zero chance the project might eventually use chips from Advanced Micro Devices instead.

That would be unacceptable. So Nvidia reportedly dropped a $150 million deposit with Crusoe and began helping recruit Meta as a tenant for the facility.

Which is one way to describe it.

Another way would be: the chip supplier is now helping finance the data center just to make sure someone shows up to buy the GPUs.

Totally normal market behavior. Nothing to see here.

The irony is that none of this was supposed to be difficult. Over the past two years the AI trade has evolved into something resembling a self-reinforcing capital spending machine. AI labs promise increasingly powerful models. Hyperscalers promise increasingly massive infrastructure. Chipmakers promise increasingly powerful hardware. Investors promise increasingly large checks.

Everyone nods along because, well, AI and shit.

Across 2024 and 2025 the numbers involved have become borderline surreal. Hyperscalers including Microsoft, Amazon, and Meta have been committing hundreds of billions in combined capital expenditures to AI infrastructure. Entire gigawatt-scale campuses are being designed around GPU clusters that can cost tens of billions of dollars before the first model ever runs.

In the past year alone we’ve seen announcements for multi-gigawatt AI campuses in Texas, Louisiana, Ohio, Indiana, and multiple locations across the Middle East. Power utilities are scrambling to keep up. Nuclear plants are being discussed as dedicated compute power sources. Some projects are literally being planned around their own on-site generation because local grids cannot handle the load.

All of this is happening on the assumption that demand for AI compute will grow exponentially and immediately. Which may very well turn out to be true.

But what the Abilene situation illustrates is that when you move from PowerPoint decks to pouring concrete, the math suddenly matters. These facilities cost tens of billions to build, require enormous amounts of electricity, and depend on financing structures that start to look a little less comfortable when interest rates are no longer pinned at zero.

That’s before you even get into the small detail that the companies renting the capacity need to actually generate revenue from all that compute.

It’s not a catastrophe. The campus is still being built. Oracle’s broader agreement to develop roughly 4.5 gigawatts of data center capacity for OpenAI remains intact, and the companies continue to pursue projects elsewhere, including a major facility near Detroit. But the symbolism is difficult to ignore, especially today.

Because the entire AI boom has been driven by the assumption that demand is so overwhelming that infrastructure can barely keep up. The idea that a hyperscale expansion might suddenly be looking for a tenant introduces a slightly less euphoric possibility: maybe the buildout is getting ahead of itself — exactly what Michael Burry suggested was happening when he compared AI to the fiber/internet boom back in December, which I pointed out.

Burry Breaks His Silence

And markets right now are not exactly in the mood for narrative disruptions.

Investors are already staring at a fairly unpleasant combination of risks. The conflict involving Iran continues to simmer with no clear path to a decisive resolution, raising the possibility of further regional escalation. Private credit — the $2 trillion corner of finance that has quietly replaced large portions of the traditional leveraged lending market — is starting to show visible signs of strain as higher rates squeeze borrowers and liquidity becomes more selective.

Overlay that with the fact that real interest rates remain positive and restrictive, meaning the cost of financing enormous infrastructure projects is no longer trivial.

Individually, none of these issues would necessarily derail markets. Together they start to look a bit like a three-headed chimera: Iran, AI capex bubble bursting and private credit vomiting up pieces of its own spleen.

For the past year markets have been happy to assume that all of this will work itself out. Now, just as investors head into a weekend already filled with nerves, one of the more exuberant pieces of the AI infrastructure narrative has quietly hit a speed bump.

And if the selling that began after that 3 PM headline is any indication, traders may spend the weekend asking an uncomfortable question: what happens if the most crowded trade in the market just developed its first real crack? You guys think about that. I’m going to happy hour.

QTR’s Disclaimer: Please read my full legal disclaimer on my About page here. This post represents my opinions only. In addition, please understand I am an idiot and often get things wrong and lose money. I may own or transact in any names mentioned in this piece at any time without warning. Contributor posts and aggregated posts have been hand selected by me, have not been fact checked and are the opinions of their authors. They are either submitted to QTR by their author, reprinted under a Creative Commons license with my best effort to uphold what the license asks, or with the permission of the author.

This is not a recommendation to buy or sell any stocks or securities, just my opinions. I often lose money on positions I trade/invest in. I may add any name mentioned in this article and sell any name mentioned in this piece at any time, without further warning. None of this is a solicitation to buy or sell securities. I may or may not own names I write about and are watching. Sometimes I’m bullish without owning things, sometimes I’m bearish and do own things. Just assume my positions could be exactly the opposite of what you think they are just in case. If I’m long I could quickly be short and vice versa. I won’t update my positions. All positions can change immediately as soon as I publish this, with or without notice and at any point I can be long, short or neutral on any position. You are on your own. Do not make decisions based on my blog. I exist on the fringe. If you see numbers and calculations of any sort, assume they are wrong and double check them. I failed Algebra in 8th grade and topped off my high school math accolades by getting a D- in remedial Calculus my senior year, before becoming an English major in college so I could bullshit my way through things easier.

The publisher does not guarantee the accuracy or completeness of the information provided in this page. These are not the opinions of any of my employers, partners, or associates. I did my best to be honest about my disclosures but can’t guarantee I am right; I write these posts after a couple beers sometimes. I edit after my posts are published because I’m impatient and lazy, so if you see a typo, check back in a half hour. Also, I just straight up get shit wrong a lot. I mention it twice because it’s that important.