Summary: Today, a look at some incredible volatility and cross-currents in the US equity market as Mag7 continues to come in for a beating as chip stocks soared once again. But then the two Korean memory makers stumble overnight, perhaps on Samsung announcing it my plow KRW 1 quadrillion (about USD 650 billion) into cap-ex over the next decade. Elsewhere, the market gave Apple the thumbs down for its greedy plan to make even more money from its customers due to higher input costs. Macro and FX and a busy event risk calendar next week and more also on today's pod, which is hosted by Saxo Global Head of Macro Strategy John J. Hardy.
The trouble with all that AI cap-ex
Gesda Capital’s Peter Garnry with a sober warning that while all of that cap-ex may be driving massive profits for some, system-wide profits in the longer run are what counts, especially whether the huge spending will generate the needed return on investment to sustain it.
Microsoft to use DeepSeek for lower cost solutions…
…as OpenAI costs mount. And OpenAI is losing tens of billions a year. Am I missing something?
FT Alphaville coverage of the CME case against the CFTC
Not to feel sorry for CME, but the regulatory regime in the US is a serious governance concern. Great coverage as nearly always.
Another one for the space- and astronomy Nerds
The ESA is launching a cool new space telescope called Plato, hopefully as early as January, that is going to take A LOT of pictures (two 81 mega-pixel cameras taking wide field views every 2.5 seconds, 24 more cameras doing so every 25 seconds - mommy I need some new multi-exabyte storage systems….) to find exoplanets that are closer to earth in size than many of the ones we have found so far. It will use the “transit method” to discover planets, in which planets traversing in front of the stars it is pointed at dim slightly as the planet transits the field of view from our vantage point. However many sun-similar stars there are with earth-similar planets orbiting at earth similar distances, we would only be able to detect 0.5% of such systems, assuming a random distribution of orbital planes. Of course, this rig will find all kinds of stuff - far more systems with close-in planets of size, etc., But it will come up with some really cool stuff to look into at unprecedented detail. ChatGPT tells me there are about 80,000 (plus-minus 20%) main-sequence stars within 10% of our Sun’s that are within the approximate 500 light-year reach of the Plato telescope’s capabilities. Throw away the binary, tertiary, etc. systems and other candidates that aren’t “just right” to study and you still probably have many star system candidates. Let’s call it the lucky 15,000 - surely, given that over 80% of stars are estimated to have planets, we’ll find some cool stuff to look into - maybe 20 or 30 with earth-like candidates we can super-zoom in on at a later date?
The company I teased on today’s SMC podcast was Blackberry, long given up as a slowly asphyxiating basket-case as its device business completely died around a decade ago and it struggled to transition to a software-only company. It now lives on mostly due to its acquisition of QNX back in 2010, a real-time operating system that is geared for security and high reliability, which has helped put it in over 275 million increasingly software-controlled vehicles - in particular for many assisted-driving features. It is also embedded in other industrial and medical devices. In addition to the increasing sophistication needs of cars and potential further growth there, the total addressable market could scale with physical AI growth if it is embedded in robots, etc. The company beat and raised with its most recent quarterly earnings report yesterday, growing top-line at 26% year-on-year. It’s a certainly a great comeback story after having fallen 98% as of late March from its 2007 high. Still, with this big advance, a lot of AI-hype and good news is already in the price - it is selling at around ten times sales and around 55 times forward earnings. I sure still miss that physical keyboard and the typing speed I could achieve on it back in the days of yore - likely no chance for a comeback there.
We’ve long discussed the lagging Mag7 on the podcast on theme of hyper-scalers (four of the seven) emptying their coffers into the hands of the chip- and other hardware makers in the ongoing epic data center buildout. But with Apple’s and even Nvidia’s recent stumbles added to the mix, the underperformance relative to the “broader” Nasdaq 100 (ugh, performance there chiefly chip-driven) has been stark in recent weeks, with a massive recent acceleration as seen in the chart below (shown as ratio of Mag7 ETF MAGS to Nasdaq 100 ETF QQQ). We have also created an admittedly cherry-picked basket of the “Nasty9” - five of the worst suffering large Software-as-a-Service makers (Adobe, ServiceNow, Salesforce, Workday and Atlassian) and four of the also-suffering and large, listed consulting services companies globally (Accenture, Cognizant Technology Solutions, Capgemini and Infosys). An evenly weighted index of the Nasty9 is indexed in the chart below to 100 as of the last trading day of 2023, more than a full year after the November 2022 introduction of ChatGPT. For maximum horror, we plot the Nasty9 to Philadelphia SOX index ratio as well. And we wonder why private equity companies are in trouble for their investments in software? Yikes.
From here, it will be a fascinating moment for the market if, or once, the hyperscalers start to reveal a flattening out, or even drop, in future cap-ex commitments on data center infrastructure. Sure, at that moment we should expect a Mag7 relative performance jump, whether relative to the broader index, but especially relative to chip stocks as the latter could suffer terrifying mark-downs. But could Mag7 stocks fall as well as the market fears that the huge spending commitments will never yield the hoped for profit margins, certainly nothing resembling legacy monopoly-driven margins? And what would be the fate for the Nasty9’s of the equity universe? Would they get a bump as the market sees a delayed “great software extinction event”? Or, do they keep heading lower still in an ongoing funeral march because AI continues to realize its potential and indeed shows steady or even accelerating progress in replacing whole categories of labor and software?
Outrageous Predictions
Chief Investment Strategist
Outrageous Predictions
Chief Investment Strategist
Outrageous Predictions
Chief Investment Strategist
Outrageous Predictions
Chief Investment Strategist
Outrageous Predictions
Global Head of Investment Strategy
Outrageous Predictions
Investor Content Strategist
Outrageous Predictions
Global Head of Macro Strategy
Outrageous Predictions
Global Head of Macro Strategy
Outrageous Predictions
Saxo Group
Outrageous Predictions
Global Head of Macro Strategy