Fx risk on Commods sort of risk on Equities? Just a complete mess

Very strange moves in everything related to public or private credit?

3 AI capex thoughts: Claude code’s potential as a genuine platform shift emboldens Big Tech 27/28 capex vs to 3m ago While it increases the usefulness of AI for corporates, it’s more impactful for smaller, more nimbler firms than larger. Latter care more about security/ permissions/ process

turning into a messy day out there for stocks perceived of being 'at risk' from AI disruption, whether valid or not. 5 of the bottom 10 performers in the Stoxx 600 today have been sold off as a result of this theme:

If we assume software doesn’t get disrupted by AI, it still leaves a very open equation of compatability. How does your agent interact with the software? Is it agent friendly? Does it have an API? What does it mean for number of seats? In some cases that’s still a big product overhaul

Apropos of nothing

Cool post

3 AI capex thoughts: Claude code’s potential as a genuine platform shift emboldens Big Tech 27/28 capex vs to 3m ago While it increases the usefulness of AI for corporates, it’s more impactful for smaller, more nimbler firms than larger. Latter care more about security/ permissions/ process

So both ZM and SKM own a decent chunk of Anthropic. Seems the best public markets way of getting exposure. Issue is the rest of business but I can see it taking off as Claude code’s popularity grows

Genuine q - why is media reporting over some this OAI stuff so bad? Both on Nvidia’s investment and OAI’s need to ASICs.

Takaichi has a habit of putting her foot in her mouth when speaking

@sellingtheta.bsky.social Any thoughts on why xau and xag riskies haven’t fallen with spot?

I'm teaching a new course at Stern on AI in Finance and opening it up! Syllabus/slides on Github: github.com/arpitrage/ai... Weekly summaries on Substack arpitrage.substack.com/p/1-three-ru... First post is on Amdahl's Law, Jevons' Paradox, and why finance was slow to learn the bitter lesson

I enjoyed the currie interview. He’s a good salesman, and wrote an extremely useful report on what he learned during his time in at GS Main lesson is commodity returns are driven by inventor flows + balances + patience Good returns from balances, great from both. Best w patience

Warsh says QE encouraged budget deficits. My experience at HM treasury in 2010 was the opposite. QE was cover for pursuing austerity

Sandisk +12%

Marginal tax rates for graduates with student loans – add interest and another 6% for postgrads. £25,000: 37% £50,270: 51% £100,000: 71%

Lower

It’s interesting how little discussion there is as to what happens to RoW equity outperformance if most people’s dollar projections come true. In 2025, h2 in Europe ex banks was mostly downgrades Japan still had material fx weakness to help support Nikkei W/ US weaker USD was helpful!

My biggest question on UK data is if a repeat of the trends in economic surprises over the last three years ocurrs UK CESI near lows also in line with previous years From a narrative perspective, post budget uncertainty could be fading, while global econ looks better. Makes long sonia dangerous

IMO the only question that matters is will Warsh be subservient to Trump? If yes, all this figuring out whether he’s hawkish or dovish, or his BS views are irrelevant

Send copper to Valhalla

Think of the narrative mismatch here when “uncertainty” is used as a driver for *checks notes* copper

But why is gold going up? It’s right here.

Given I worked till 5am yday, safe to say I don’t care about fed tonight

he's not here to close spreads

Most my convos this week w clients were about if UST were signalling desire for broader USD weakness. I felt no, but this will absolutely stoke those fears

Booom But seriously, while they are smashing it, how can you have any confidence in their moat?

When America’s biggest Shinzo Abe fan checks rates to try to call a top in USDJPY shortly before appointing a possible Kuroda clone to run the Fed, you should probably listen?

Copper needs to trade 15% cheaper imo

So in the AI space, every one of these price gains will have to be eaten by Hyperscalers right? Nvidia isn’t lowering gross margins yet - that’s a MT/ LT story So for now, higher memory, copper, powder costs are all reducing ROI

Right up my alley

Quick shout out to @tobiasharris.bsky.social for absolutely sterling work covering Japan’s complex (and increasingly fast-moving) politics. His newsletter is an invaluable resource for anyone working/studying/investing in the country.

My note from Tuesday

I The way the Treasury has historically intervened is via their Exchange Rate Stabilisation Fund. This Fund is currently about $200bn so much smaller than FX reserves held by MoF.

UST rate check all but confirmed… truly back to the 90’s type stuff.

Rejoined the Macro Hive podcast to speak with @vkmacro.bsky.social about what the heck is going on in the oil market and how to parse the numerous conflicting data points/narratives. Viresh is one of the best cross-asset analysts—also superb oil-specific insights!—with whom I regularly chat.

Meanwhile, UST 5s continues its cheapening trend YTD vs broad curve. Typical price action as cut cycle nears an end 2s5s30s curve

Absolute 🔥 by Ben Thompson What could really kill the AI bubble is that not enough people were investing in capacity in 2025, meaning there’s a massive shortage of compute in 2028/29 Not that we invested too much.

Ben Bajarin: we cannot trace grace Blackwell GPU pricing since you have to be an established corporation or startup and be willing to engage in a long term contract with a hyperscaler to even get a quote

20y fwd 10y JGB now trades above 20y10y UST

Everything’s looking more pro growth isn’t it?

I am enjoying this discourse: Q: what is nvda’s moat? A: software w/ cuda Q: What is AI doing? A: Reducing value of software as it can be replaced easily and also customised better Q: But what does that mean for NVDA’s most? A: …

Always a fun read

Anyone have good tips on sell discipline during parabolas?

Thread on some work I’ve done in PE

Sorry I’m not spending even 1m watching any of this. Only thing I’ve watched so far is Carney and AI stuff

Over the past three months foreign investors have net taken money out of SK equities. Since the start of 2025, they’ve not invested net any new cash. Definitely some profit taking, but shows that no one believes the AI/ memory rally. How do you expect the locals to?

Know what you own and why you own it. But actually it’s interesting how poor the S&P broad value index is given their loose definitions. I’m fine with meta being in a value index.

Wages rolling over, unemployment grinding up, job losses continuing, an administration prices are likely less of an issue this year. Taylor thinks neutral is around 2.75%. I’m not saying I know where it is, but does open up room for rates to move lower.