One of my main reasons for writing the January piece was to force the book into the open. I've delayed it, which is a tell. Writing theses is cheap. Writing what you actually own is a commitment, and I am trying to stay intellectually flexible, not married to a set of positions.
So, still not a holdings dump. No tickers, no weights. That would be unwise and slightly beside the point. The point is how the book is built, and why I haven't flipped it because a jackson hole speech made the front-end twitch.
High-level: I'm still risk-on.
"Investing is easy in bull markets" is a truism. I don't blindly or naively expect a march upwards. Friday was a useful reminder. Warsh stood up at Jackson Hole, said that unless inflation is moving to target "clearly and at sufficient speed," the Fed still has "work to do," and the market did what it always does when a new Fed chair decides to sound like a central banker: two-year yields jumped, hike odds for September went from a third to a coin-flip, chips wobbled, Nvidia gave back a slab of the previous session's joy. The S&P sat around 7,700 and pretended this was drama.
It's noise around the arc. The arc hasn't changed.
Current construction:
I'm most comfortable in what many still see as highly speculative assets, in the themes I sketched in January, and mostly in thematic ETFs rather than individual stocks.
That last bit matters more than the first. Single names are a personality test. Baskets are an admission that I'm trying to own a future, not win a stock-picking contest. The Mag7 can be eldritch entities. I don't need to decide which one is Cthulhu.
- We are still hurtling towards a sci-fi future, so the book is long that complex. Space, robotics, drones, autonomous vehicles... and of course AI, which is both separate from and related to the above. I want a diversified basket of speculative frontiers, not a house view on which lab ships the next model. Culture novels, Hyperion, Deus Ex, All Of the Above. Thematic vehicles are the cleanest way to stay long that without pretending I can underwrite every name.
- Real resources were always the constraint, and they still are. Energy, defence, infrastructure, the stuff that has to exist in the world before the software can. Resource-competition wars are not a sidebar to the bond market. They're part of why the old yield-suppression machine is dying. Within that, solar is at a sweet spot of underloved (unnecessary concerns around rightwing scepticism, and Chinese overproduction) and sci-fi tailwinds (the Sun as the rightful King among energy sources in the Solar System). Nuclear is the crowded AI-power trade. Solar is still the one people want to argue with. I like that.
- Almost no duration as a core holding. A long-end call is a sensible, cheap-ish option. It is not a high-conviction call. I'm not pounding the table on it, and I'm not going to be the person who heroically longs the long end because the yield "looks high." For a retail-shaped book that can be as crude as deep OTM calls on ultra long G10 duration ETFs. Convexity on a policy accident. Not a bond portfolio.
- Crypto stays, sized like a thesis, not a religion. Quantum concern is still, imo, a real factor in bitcoin's lag. People will have extremely strong views on whether this is right or wrong. I'm just confident it still matters.
What I'm not doing: trying to be clever in European equities as an identity. For many in Europe it's fashionable to dislike both the US market and "AI" stocks. This is based on a hope it all comes crumbling down — copium. US capital markets are winner-takes-all for the rest of the Western world. A threat to this is some form of capital control framework. Trump decides RoW can't own US equities? Non-zero probability, but I'm currently investing firmly on the premise this isn't happening.
Property investing is over for retail. Complicated topic, but plausibly a new source of structural demand for equities as mom & pop landlords realise the jig is up, and VT & chill is superior return (and return/stress ratio). The "VT and chill" crew are right, for the wrong reasons. It is not a good investment because we live in an EMH NG++ world, but because there is a socio-political imperative to keep the SPX marching higher. My book is a levered, weirder expression of that same bet, not a rejection of it.
The rates thing (not a duration call):
The option in (3) is sitting on top of something bigger than Warsh practising hawkishness for a weekend.
There's a regime on in rates that is the exact inverse of what my generation was trained on after the financial crisis. Post-GFC, the sell side spent a decade calling yields higher, higher, higher. They were consistently too high. The trade, the career, the muscle memory, was to fade that and look for lower. QE made a whole cohort of PMs into lower-yield animals.
Now the sell side has flipped. Analysts treat yields as too high, and are constantly forecasting a rally. Completely inverse of the previous period. That's not a small colour comment. It's a reversion to the classic university textbook model of bonds — term premia, fiscal arithmetic, real-resource constraints — and it's throwing off people of my generation who are still scanning the tape for lower and lower yields.
Resource competition, defence build, energy, the whole physical-constraint stack: that's the fundamental, not a cycle. Central banks are less able to run the bulky QE-era playbook over the top of it. The expenses are real. The supply of paper is real.
Europe is where this spills. France is the one. Some of the European sovereign complex is actually at risk from this change in the global rates regime, not from one messy budget week. AI belongs in that story too. The credit impulse that used to come off a high-skill labour / income base is shifting, and it's hitting Europe's overall fiscal and financial conditions. Fundamental change, not a hawkish speech.
Nobody is digesting 5% gilts. Very few people are reaching for the 4s either. France is where you'd actually be interested, as a concern, not as a bid.
And this is not "SpaceX doesn't need capital." They will raise again. They already did: equity in June, then a slug of senior notes a fortnight later in the mid-5s to high-6s. When a name like that borrows at 6 or 7, you can reasonably forecast that returns on capital far exceed the coupon. That is why the paper clears. The UK sovereign at 5 is a completely different kind of maths. There is no equivalent ROC story. Private capital is not going to sit in gilts at 5, or French 4s, when the sci-fi complex can print 6-handle debt against a return that, at least on a reasonable forecast, still covers it by a mile.
I flagged this in January. Some corporates may compete with nation states literally as well as metaphorically. The long-term version is two tech trees. The short-term version is more vulgar: not that they self-finance, but that they can keep coming to the market, at coupons a sovereign would kill for in real terms, because the return on capital is a different object to "the UK can roll its debt."
Gilts have been the ugly child of the G10, 10y through 5%. JGBs have done the thing I flagged in January: Japan is no longer a museum of zero. 10y toward 3%, 30s through 4%. France still looks like a university textbook. German 10s are still around 3% and still not a destination. US 10y around 4.7% and the 30y around 5.2% after Warsh is not a gift. It's the market starting to price that debt and deficit actually matter.
QE → QT only → QE again when CBs realise the horror of what they are doing. That's what the duration option is for. It is not a view that we have a great long-end entry. It's a cheap ticket on the accident. I reserve the right to be wrong about the timing, obviously.
What would change the book:
Being ideologically inflexible is an awful way to invest. So the actual kill-switches, not the vibes:
- A genuine capital-control regime that breaks Western ownership of US risk assets. I'm not positioned for that. If it starts looking real, the book has to change, fast.
- Evidence that the sci-fi complex is a capex dead-end rather than a noisy adoption curve. A "bad" AI event or a drone attack on a Western icon would, perversely, not be enough on its own. I already expect something in that family. What would matter is if the political response is strangulation rather than integration.
- A bond market accident that is actually about solvency, not about Warsh practising hawkishness. Japan and France are the obvious candidates. If that accident comes, the duration option is supposed to pay. If it comes and the option was too cute to hold, that's on me. I'm not pretending I have the timing.
- Climate / energy policy that makes solar politically unsellable in the West for a decade. I don't think that's the path. Adaptation, not stalling. But I can be wrong, and nuclear-as-AI-power can crowd everything else out for longer than is comfortable.
I will not change the book because semiconductors had a bad Friday, or because someone on the timeline declared the AI boom over for the fifth time this year.
I'm risk-on because that's where I'm most comfortable, not because I have a model that says SPX 10,000. The lanterns still point at the same corridor. I'm experimenting with the optimal format right now, so watch this space.
Views are my own. This is not investment advice. This venture is personal and unaffiliated with any professional role.