(Part 2 of 5) Midyear State of Data Centers and the AI Bubble Report - Pt. 2 Data Centers and Data Center Construction

In Part 2 of 5 of our Midyear State of Data Centers and the AI Bubble - we discuss the public equities side of the trade, the financing wobbles, are data centers shale on steroids?, and much more...

Don Johnson (@DonMiami3), Chief Economist

Good Friday evening MacroEdge Readers and Community,

This evening we’re going to have a shorter (following some feedback and comments to Part 1) look at the state of data centers and data center construction. In Part 1, I focused heavily on the what, why, and buildout - so if you haven’t yet read through Part 1, I recommend doing so that Part 2 makes sense.

Part 1 Table of Contents:

The report came at an apt time in the momentum/AI/data center trade - as large unwinds began to occur over the last 14 trading days (especially with the liquidation of the ‘Situational Awareness’ public equities book). While many are choosing to cite their elevated Anthropic pre-IPO buy-in as evidence ‘still’ of Aschenbrenner being a genius - I suspect that over time that label will wane, and eventually the mask will be lifted. Every cycle, these false market oracles are created - and I don’t think that this one is done yet, especially judging by South Korean intervention in markets - and massive intervention in Japanese currency markets. While somewhat unrelated to what we’ll be discussing tonight, the global markets are essentially all one trade now. For the better part of a year now, I’ve warned that correlations continue to grow across this entire trade - and as the financing of it gets more interconnected and government-backed, those correlations will only grow further.

Look at the US AI ETF (AIS) versus the Nikkei 225, as one example that I’ve frequently highlighted:

The KOSPI correlation is even stronger:

Well… what does this have to do with data centers? A whole lot. The public equity bubble, as we discuss below, has been a huge enabler of the entire data center boom, and as it matures and eventually boosts, we’ll see a shift toward government-financing through the military, etc as the private sector math stops penciling. One of the greatest risks to the entire buildout with future construction is now opposition at the local level to the data center facilities through the withdrawal of local tax incentives. This is a strategy already proving successful in Texas - and even though the massive buildout is well underway and (in my opinion) shifting toward that mature phase for the current cycle - there is now downside room on the construction side to create an air pocket before we start to see intervention on the construction side more directly. Data centers have been utilized like a Soviet Union-style economic program to prop up, stimulate, and ‘grow’ the economy - and they’ve done a masterful job of keeping up construction employment that way too. Look at the latest construction-related data via FRED:

(Total Construction Spend, Nominal):

(YoY - % Terms Basis)

Ex-res real estate:

Is there any public employment data to be trusted anymore? It’s certainly getting harder:

Specialty contractor employment:

While I could claim that I don’t have a clue why we’re undergoing a buildout of this scale for something that is unlikely to ever generate long-term profits in its current form - I do have many clues - and the answers are not pretty. This entire buildout has served as a beautiful faux growth source during a period of innovation stagnation - especially at the ‘hyperscalers’ and Magnificent 7 companies. It’s a question that I pose frequently to my community on X - in the last 5–10 years, from a technology standpoint (outside of LLMs - which are also questionably labelled AI in their current form) that has revolutionized anything? Not particularly. Peak innovation comes at peak cycles, and we’re far more interested today in a Roman-style print-and-forget philosophy until something breaks - and the bond market is finally starting to reprice that reality - which will also inevitably negatively impact the entire data center ecosystem.

Without rambling on too much more to finish off the introduction - we’re going to jump right into the content we’ll be covering tonight - which is focusing not on the background, tax incentives, buildouts, construction data, etc. - but on the following:

• (Part 2) The Public Equities

• (Part 2) The Financing Wobbles

• (Part 2) Are Data Centers Shale on Steroids?

• (Part 2) Data Centers and the Rings of Empty Chinese Cities

• (Part 2) Finishing Comments - A National Security Pivot

Let’s dive in.

Catch up on Part 1 of 5 - Midyear State of Data Centers and the AI Bubble Report

Next in the Series and Upcoming Reports

In the ‘Midyear State of Data Centers and the AI Bubble Report’ - we are covering everything over five parts. Below are following three parts that are still to come (and will be released on Wednesday for the next three weeks):

• Part 3: A Look at the LLMs and Assessing the Moats

• Part 4: How Big is the Bubble, and Will There be Bailouts?

• Part 5: A Summary and What’s Ahead for the Second Half

Outside of this series - over the weekend - we’re going to be covering a lot of ground. In addition to this Part 2 report, we’ll also have:

• (Saturday) Redeye Macro Note

• (Sunday) Weekly Macro Note: The Zimbabwe Roadmap

• (Tuesday) War Note and What’s Coming Soon

• (Wednesday) Product/Offering Previews, Partnerships, and an Update

• And more on the way

Not yet a MacroEdge Ozone paid subscriber? Upgrade to paid below to get all of our research, data, portfolio strategy, commentary, and much more:

The Public Equities

Let’s talk about the pure-play US equities that have been riding shotgun on this entire data center buildout, because the divergence between them tells you almost as much as the spending and construction employment data we just went through (not very much)...

Continue reading below with Ozone <The Public Equities continued, The Financing Wobbles, Are Data Centers Shale on Steroids?, & much more…>

Upgrade below: