14 charts riffing on themes

Healthcare angels; AI Capex; the real manufacturing boom; healthcare-foreign-born-job-maker 'in balance'; no immigrants, no blackjack; the last pro-cycle fades

• fallen healthcare angels;

• AI Capex;

• the real manufacturing boom;

• healthcare-foreign-born-job-maker 'in balance';

• no immigrants, no blackjack;

• the last pro-cycle fades

---

👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. 

👋👋👋Random Walk has been piloting some other initiatives and now would like to hear from broader universe of you:

(1) 🛎️ Schedule a time to chat with me. I want to know what would be valuable to you. 

(2) 💡 Find out more about Random Walk Idea Dinners. High-Signal Serendipity.

---

Fallen healthcare angels (reprise)

Just another chart to illustrate the earlier point about the insurance bullwhip of costs preceding revenues:

The Medicare Cost index jumped, without a corresponding lift in consumer healthcare services spending.

In other words, costs increase for the insurers first and then those costs are passed on to the consumers (in the form of rising PCE, just around the corner).

Anyways, UNH 0.00%↑ already had a win yesterday, so best to be lucky and right.

---

ICYMI

Of profitability

For VC, the song remains the same, which is why it's totally different now

Silver tsunami, reshoring, and the unbearable lightness of tariffs

10 charts on Doing More with Less

Data-driven real estate, with Lance Lambert of ResiClub Analytics

---

AI Capex is the Everything Cycle

Random Walk has been on a Cembalest kick, but he keeps coming out with bangers (albeit, buried bangers in his longer manifestos).

This is just a great visual on the impact of AI Capex on the overall economy, right now:

Eye on the Market

AI Capex contributes just under 50% of GDP growth.

Pretty incredible.

Here’s another cut for a sense of proportion on all this AI investment, albeit focused specifically on the real estate (aka Data Centers).

Data center spending alone is on pace to eclipse the total amount spent on general office construction:

Office and Data Center spending looks like it will converge at ~$45B sometime in the next year.

Just three years ago, Office and Data Centers were separated by ~$60B.

And here’s yet another chart for scale, although this one has a bit of a surprise:

Wells Fargo

Investment in data centers is on the verge of surpassing the previous industrial leaders, Chemical and Other . . . but data centers got nothing on Computer/Electric.

Manufacturing construction spend on “Computer/Electric” (i.e. semis, and related tech hardware) went from basically nothing to ~$130B in ~2 years. That’s 3x what’s being spent on data center construction right now.

I genuinely had no idea we had started to spend quite so much money on semis and other tech hardware stuff. The Inflation Non-Reduction Act was a helluva stimulant, no doubt.

In all events, idk if/when we’ll see big productivity gains—I do think we will, eventually—but if not, it won’t be for lack of trying, that’s for sure.¹

---

Healthcare makes all the jobs, job-growth is a foreign born story (reprise)

Another day, another Random Walk-was-totally-right-all-along article about how healthcare makes all the jobs, and wouldn’t you know it, but it turns out job growth was a foreign born story.

Oh no, why did tariffs do this to us:

BanCreek

After the rest of the job market mostly recovered its pandemic losses, it’s been healthcare all the way up (down?).

You see, if you were under the impression that the labor market was “strong,” that’s because you were told that labor market catch-up/recovery was labor market “growth:”

Yardeni

The supply of workers dropped, while consumption/demand surged, but it still took ~3-4 years for the labor market to finally recover its trend.

In the intervening period of the “tight but not strong” job market, wages—and then prices—blew out the wazoo.

But now, we’re pretty clearly back in balance:

Wage-growth for job stayers crept ahead of job-switchers, which is about a strong a sign of balanced supply-demand for people, as you’re gonna find.

Incidentally, the last time stayers ran ahead of switchers was the post-GFC period, notably a very weak period for labor demand.

For now though, it’s mostly job market stasis. No hiring, firing, or quitting.

By yet another measure, people really are just staying put, although ever so slightly less:

BofA

The “Payroll Disruption Rate” (“PDR”) crept up in July, but is still lower than it was a year ago.

Payroll can be disrupted by (a) taking leave; (b) getting a new job; or (c) getting fired. In this case, to the extent PDR is meaningful at all, my guess it’s (c) driving most of the change, but who knows, maybe it’s positive jolt of labor market dynamism.

Anyways, back to the story, the question is now (as it was then): once healthcare is all caught up, what happens next?²

We don’t really know, yet. There is AI, the job-maker (not job-taker). There is, perhaps, reshoring, with smaller, less-profitable, but more people-intensive firms. But mostly, we just don’t know, and for now, it’s just healthcare and AI Capex that’s keeping things afloat.

---

Rotation to the National Nursing Home is Running Out of Foreign Born Workers (reprise)

As a final aside, even if we did somehow create all kinds of new labor demand, it’s unclear that we’d necessarily have the bodies to fill the roles.

• We know that migrants were critical to closing the “labor shortage,” because the Great Replacement is real, deflationary, and pro-growth.

• We don’t know if more migrants would create more jobs (now that things are in-balance), or if more migrants would create more unemployment (for now), like in Canada.

What we do kinda know is that job growth for “foreign born” workers has tapered:

Employ America

Sectors with a higher concentration of “undocumented” or illegal immigrants have experienced a decrease in employment-growth.³

Now, is that because the supply of foreign born workers has dried up, or is it because demand for, e.g. construction and retail/hospitality has gone flat (or both)?

Or maybe it’s this:

Twitter

Automation for the win?

Again, we don’t really know yet, but we certainly do know that construction—the only other pro-cyclical driver outside of AI—is trending the wrong way:

Wells Fargo

Declines in housing starts tend to lead declines in construction employment.

It’s a real shocker, but when there is no housing shortage, and not much population growth, there is much less reason to build new houses.

---

Previously, on Random Walk

---

Random Walk is an idea company dedicated to the discovery of idea alpha. Find differentiated data, perspectives and people, and keep your information mix lively. A foolish consistency is the hobgoblin of small minds. Fight the Great Idea Stagnation. Join Random Walk. Follow me on twitter. Follow me on substack:

1. I also continue to think that productivity gains will show up in unexpected ways. Just another example, the sheer power requirements demanded by AI compute are an engineering challenge, in and of themselves:

The power-density requirements for AI servers are expected to triple in the next ~2 years. I know basically nothing about electrical engineering, but it seems to me that by opening up a whole other innovation vector, we’re bound to discover useful things along the way (that have nothing to do with AI).

2. Hilariously, healthcare is also a non-trivial driver of e-commerce gains

Consumer Edge

Telehealth and Pharma atop the global ecommerce leaderboard.

3. Illegal immigrant is still the correct, statutory term. I find hand-wavey euphemisms irksome, all the more so when they become socially enforced taboos.