Outrageous Predictions
Carry trade unwind brings USD/JPY to 100 and Japan’s next asset bubble
Charu Chanana
Chief Investment Strategist
Summary: Today, we discuss the ongoing "AI Overlay" trade, note another AI-related company that is not a hyperscaler, but is set to spend up to USD 200 billion on capacity expansions in coming years. Elsewhere, we discuss the strength of US treasury and Japans' government bond markets and whether this is contributing to pressure on precious metals. As well, we ponder whether both the US dollar and yen might strengthen against the other major currencies and the next keys for sterling direction. Today's pod features Saxo Head of Commodity Strategy Ole Hansen and is hosted by Saxo Global Head of Macro Strategy John J. Hardy.
Weak companies for the win - an AI story? Goldman Sachs and its (dodgy?) long/short AI winners/losers basket. Getting into the AI weeds can be enlightening Craig Tindale weighs in against the very high cost of “free markets” The original sin of post-WWII US Policy: making the USD the reserve currency. Mike Green with a thoughtful piece on what inequality really means for outcomes That’s a lot of cheddar! Micron is set to spend USD 200 billion on expanding memory chip production, including the largest cap-ex in New York State’s history with a USD 100 billion factory there. The company expects is will have tripled sales in just two years from its quarter ending August 2024 to the quarter ending this August. More importantly, profits are exploding due to pricing power amidst a shortage of memory modules relative to demand linked largely to AI data centers. Throughout its history, the company has seen wild swings in profits, with more negative earnings years than positive ones from 2001 through 2012, for example. It’s best ever year was 2022 (through end August), when it earned just short of USD 9 billion. This year through August it is expected to earning USD 38.7 billion and next year over USD 50 billion. Still, once a hardware company gets to such a size, it is risky to project forward demand will continue to achieve a strong growth trajectory…
Andreas Steno Larsen out asking whether what the market is telling us is that a wide swathe of companies, especially many with low margins that aren’t even investing in AI, may be set to reap its benefits and enjoy a productivity boost and margin expansion. It’s as good an explainer as any for why many weak companies have outperformed recently, but is it a narrative flavor of the day or durable? I’m not convinced of the latter.
I pass this along only because I promised to in the podcast. I am somewhat skeptical although for a price, I am sure Goldman can provide you with a very long research piece on why it chose as it did. Some of the names on each side make sense, others don’t.
I have only listened to a fraction of this interview on the Dwarkesh Podcast with Anthropic CEO Dario Amodei, but it already drove me to do some further research that is enlightening for understanding the AI phenomenon, especially why many of the the major players insist on engaging in the brute computing force arms race, as many are convinced that scaling is the only way to reap AI’s gains. There is a built-in irony here - if everyone has access to the same tools, the biggest data center “wins”, but what does winning look like if the productivity gains aren’t there to pay for the effort and especially if a world model approach disrupts the brute force approach?
Craig Tindale out with another sharp piece on the very high cost to national security that accumulated during the decades of “free markets” : “A nation that permits a strategic rival to produce its core materials in exchange for marginal price relief is engaging in suicidal stupidity masked as intellectual sophistication. It is pawning its industrial base for a quarterly report.” Ouch.
A LinkedIn post from the iconoclastic Steve Keen who suggests that the current geopolitical backdrop was pre-determined the moment that the US made its own currency the global reserve currency rather than Keynes’ Bancor. Key points in this - even I highly doubt that the rational path he recommends is the one the powers that be eventually take, however....
The implication is that excessive inequality is very bad. I always enjoy the way Mike Green thinks and makes his case, almost no matter the topic. Here, he takes a quantitative approach to illustrating how wealth inequality drives very different outcomes for people of equal talent/intelligence.Chart of the Day - Micron (MU)