5 Idea Friday: capex, slow n' steady (and others)
The surprisingly linear ascent of tech capex, mysteries of biotech capital allocation, ex-chinese etailers, never go full French, and CRE's bottom (and the niche that could)
tech capex, less cyclical?
Random Walk ponders capital allocation in biotech (and some weirdness)
China levels the playing field for US etailers(?)
Never go Full French (pension)
CRE’s bottom, and the little niche that could
👉👉👉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.Five Idea Friday
Five little open-ended ditties to carry you through the weekend.
Tech capex, slow and steady
When it comes to fixed investment in the technology sector, I was surprised to learn that growth has been fairly steady:
Fixed-investment in High Tech has grown mostly linearly for ~65 years.
Obviously, as of recently, tech capex has inflected upwards (as it did in the lead-up to the Dotcom Bubble, and in the late 70s, when semis first took off), but over a longer time horizon, capex has been very even-keeled.1 It demonstrates very little of the spikey cyclicality of either Residential investment, or the other categories of non-residential investment (e.g. Office, industrial, utilities, etc.).
I’m not really sure what to make of it, only to note that I wasn’t really expecting it, because big tech capex is generally considered pretty cyclical.
I suppose one takeaway is that tech capex cycles are different than broader economic cycles.
Unlike the latter, which run more or less with the broader economy, tech capex runs closer in-line with technical breakthroughs (and therefore cyclicality is both uncorrelated to everything else, and relatively muted, in the bigger scheme of things).
The other thing, I suppose, is that tech is its own customer, in the sense of tech creates its own demand for more tech (unlike, say, houses or office buildings, which need more people and/or firms to drive demand).
More and better semis, begets more and better software, which more and better semis, all of which eventually begets cloud, which also begets more and better software (and more data), which begets more cloud, and onwards and upwards. That helps explain why tech investment as a % of GDP continues to rise (because the tech sector, gets bigger and bigger).
All of this is to say, is that it takes some of the edge off the recent capex inflection.
You can’t really “overbuild” for tech in the same way that you can “overbuild” for people, insofar as there is no comparable outer limit on how much compute (or data) software can consume.2 Consumers only need so much food, shelter, and stuff . . . but software can keep getting that much faster, stronger, and more capable.
Maybe. I guess we’ll find out.
ICYMI
How exactly does biotech investment work?
Random Walk doesn’t really know much about biotech, but I’m starting to learn.
The first (running) observation is that China is Eating Biotech, and that still appears to be the case.
Just the latest in an increasing line of chart-samples is China’s growing share of, well, everything biotech related:
China has more publications, more IPOs, more pipelines, and more licensing deals with big US/Europe pharma than they did before.
That’s interesting so far as it goes, but I’m starting to get the sense that there’s some capital allocation weirdness afoot. Weird to me, at least, but I’m sure that someone more knowledgeable can put a finer point on it.
Consider the current (as of last year) state of biotech solvency:
Nearly 40% of biotechs have (or had) less than a year of cash left, which is worse than the boom years, but even worse than 2019.
If nearly half of biotechs are on the path to insolvency (absent some intervention), then we’re going to have a hard time keeping whatever lead we had on China. But also, it seems pretty clear (because there’s no shortage of capital generally) that investors are not seeing the risk-return in funding biotech (for whatever reason).
Another interesting thing is that, counter-intuitively, big pharma R&D spending has actually increased (relative to the pre-pandemic prior):
31% of big pharma capital allocation is going to R&D, which is higher than every year, but peak ZIRP.
Likewise, dividends and buybacks (as a share of capital allocation) are smaller than every year but 2020.
So, it’s not like the industry is just sitting on its thumbs, getting fat on cash. Big Pharma is well-aware of the coming “Growth Gap” (i.e. the expiration of lucrative patents) that starts to kick-in in earnest ~2027. They appear to have updated their investment decisions accordingly, although, I suppose not so much that it would put money in the coffers of biotechs writ-large.
Again, I don’t know what this all means (just yet), but I find both observations striking, and their juxtaposition striking, a well. Is Big Pharma going-it alone? Is relief around the corner? Is there some other reason that investment appears to be concentrating? Idk.
Chines e-tailers become ex-Chinese e-tailers
I suppose I’ve been on something of a China-kick, but it’s not intentional.
I just found this mostly amusing, but it also speaks to some of the difficulty of designing effective “trade policy.”
When China announced that it was going to start taxing Chinese e-tailers (or rather, start collecting the data to make taxation possible), lots of e-tailers suddenly moved their domicile elsewhere:
That’s ~1,500 ex-Chinese etailers in a period of ~3 months.
While still representing less than 1% of the roughly 300,000 Chinese sellers active on Amazon.com, the accelerating trend signals a fundamental shift in regulatory pressure. The catalyst is almost certainly China’s State Council Order No. 810 . . . which compelled platforms, including Amazon, to submit seller identity, transaction, revenue, and fee data quarterly. The first mandatory submission covering third-quarter data was due October 31, 2025 – precisely when the exodus accelerated.
Hong Kong’s appeal is straightforward. Where mainland China taxes resident enterprises on worldwide income at a 25% corporate rate, Hong Kong offers a territorial system that taxes only locally sourced profits at 8.25% on the first HKD 2 million, then 16.5% thereafter – with no VAT, no sales tax, and no capital gains tax.
For cross-border sellers, the differential is substantial. According to Marketplace Pulse estimates, 75% of relocating sellers generate less than $1 million annually, with 19% in the $1-5 million range – precisely the cohort most vulnerable to margin compression from sudden tax enforcement.
So, it’s only ~1% of the Chinese seller-verse (by count), but at least for smaller shops, the tax arbitrage is (and has been) apparently a pretty big deal.
It’s a funny thing too, considering China’s broader efforts to support its exports. It’s American sellers who have often complained of an unlevel playing field, and now China is (apparently) stepping up to level the playing field, at least somewhat. I guess those fiscal pressures are adding up?
Or maybe it’s just a mistake to assume that China (or any country) operates with a single-mind, and/or coherent policy-set.
As an aside, Bloomberg did a scrolly-tell of some of the price and margin pressure that Chinese companies are facing, as they systematically attempt to sell all the goods, at the same time as the US tries to throw sand-in-the-gears:
China is (alone) under deflationary pressure:
Chinese firms are increasingly heading towards insolvency, cutting staff, and reducing spend:
Loss-makers are on the rise, and even the survivors, are reporting reduced margins:
So no, China is not going to “win” the trade war (even if that might be somewhat besides the point). The trade war hits China very much where it hurts.
Never go full-french
I’ve shared a version of this before, but it’s still so-striking, I feel compelled to share it again.
In France, pensioners make more money than the average working age adult:
Pension income in France is ~102% of the working folks.
After France, Italy is running a close second, and the US is pretty high, as well.
That’s just bonkers. If that doesn’t illustrate some of the time-preference vulnerabilities of democratic regimes, then idk what does. Obviously it’s not sustainable, but it’s also just shameful.
CRE hits bottom?
And finally, after years of firesales and markdowns (because that’s what happens when you don’t have a 30-year fixed rate mortgage), it appears that maybe CRE is turning a corner?
Buying outpaced selling for the first time since rates went up:
Institutions were net-buyers of real estate by a bare $4.6B in 2025.
Huzzah! And rates are still pretty high.
To be clear, overall transaction volume is still pretty low.
Likewise, most RE asset values either underwater, or barely above water (and dramatically so, off their ZIRP-peaks):
The overall commercial property index (as measured by Green St.) is still below the pre-pandemic level.
The three out-performers are industrial (i.e. data centers), self-storage, and the hottest niche only some folks have heard of: manufactured homes.
It’s a good time to be in the trailer-park business, I guess. But maybe the rest of the real estate industry will get some wind its sails soon, too.
Previously, on Random Walk
Private Credit and Insurance, two peas in a pod (reprise), and a chart dump on default rates
five charts on the rise of private credit in life insurance
Energy in 1776
It’s July 4th, so Happy Birthday America, and we’re going to keep it light and only semi-topical.
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:
As per the BEA, High Tech includes computers, communication equipment, software and R&D, and semis.
Obviously, you can “overbuild” for tech (e.g. too many smartphone chips without people to buy new smartphones), but it’s more of a timing thing—the infrastructure investments pre-Dotcom weren’t wrong, they were early.



























"Chinese etailers, uh find a way."