Weekly Wrap – The Trillion Dollar Question... THIS MARKET!
Plus an important note for premium subscribers.
▶A quickie today, but first if you’re a paid/premium subscriber: If you did not receive yesterday’s full report headlined, “Why CoreWeave is Worth Watching” – or only received a preview, with a link to the full report – please let me know by responding to this email. I know that some of you didn’t. If you missed it entirely , here’s a quick summary...
I’m not piling into the CoreWeave debate, but I believe the company’s rollout of its CoreWeave Ventures is one of those events you’ll look back on when you try to connect the dots in this ever expanding everything, everywhere, all-at-once gold rush into AI infrastructure. So will the role of private credit. As I say in this report, “The only thing more dangerous than saying it’s different this time is thinking this time will be the same as the last.” I get it, but the parallels with prior cycles are eerily and strikingly similar... down to the hubris.
It’s quite a story. If you get a chance to read it, I think you will enjoy it.
▶As expected, I received some pushback, including this from a friend who has been early and remains very long on the AI play... and in addition to Nvidia and others, is long CoreWeave $CRWV: “A lot of people are drawing the comparison [as what happened in the telecom/Internet infrastructure build-out], but I also think it’s important to ask what is different about both periods and the Capex being spent.”
As someone who was short many of the fiber companies in the telecom bubble, he went on to say that one big difference was that overcapacity of fiber was more likely than the buildout of AI infrastructure – at least right now – because fiber would last for decades. Back then, he recalled, networking switches were installed “when demand for them didn’t exist” for the sake of meeting the needs of video on demand via the Internet, “which took years to arrive as a technology.” To be sure, at the time Netflix was still 100% DVDs as cable’s video on demand seemed the most promising.
By contrast, he says, “current chips being installed are not nearly obsolete, but will be downgraded, for the use of so-called inference, after two years of work.” In the world of AI, inference refers to the AI making predictions and decisions... and, yes, that includes some of those absolutisms that can wind up being hallucinations. My pal’s point was that those chips will be replaced by higher performance chips... and as a result, the big hyperscalers like Meta, Google and Microsoft are spreading out demand over years resulting in the ongoing need for more capacity. “So,” he says, “I think it’s a fun comparison, and will keep us climbing the wall of worry, but these are two different types of cycles.”
▶Speaking of walls of worry... How long this one lasts – that’s the trillion dollar question. It’s what everybody is talking about, especially with the gains some stocks are making on a daily basis. That kind of action simply is not normal. To that point: A few weeks ago I quoted Steve Straza of All Star Charts in my “Throw-a-Dart Market” report as saying “the worst ones” are starting to rise. By “the worst ones” he was referring to what technical analysts call “breadth expansion.” Or in plain English: “They are buying everything. So buying lows is actually working better than buying highs. For traders. And that tends to happen in later innings of bull cycles.” He’s not alone with those thoughts...t
Another friend and long-time subscriber, who is long-biased and does intense fundamental work, also happens to be a long-time user of technical analysis, with market psychology as an overlay. He says it has helped him over the decades. In trying to make sense of today’s market, he harkens back to a quote from the legendary and late Richard Russell, who wrote the Dow Theory newsletter. In 2009, when gold had been spiraling higher in what surely seemed like a bubble, Russell wrote that “the first and second psychological phases of gold are over.” He then added...
I have never seen a bull market of this size end without a highly-speculative third phase explosion. What we see on the chart over the last five years is a huge accumulation pattern in the shape of a head-and-shoulders bottom. The breakout would or will come if gold can climb above 1000 and then to a new high above 1005. Somewhere ahead I expect to see a worldwide panic – scramble for gold as it dawns on the world population that they have been hoodwinked by the central banks’ creation of so-called paper wealth.
Gold is nearly 4-times that now, and we’re still debating the wisdom of the Fed. Through it all, my friend – who lives in the upper Midwest and manages his own money and is extraordinarily independent and thoughtful – used Russell’s words as a springboard to do his own research into long bull markets that in an explosive phase. In a private report he wrote in 2011, he went back to 1986 and studied multiple bubbles since then. His analysis showed that these blowoffs lasted between four and six months. That was a different time, of course, pre-pod shop and before all-things-algos ruled the markets. But it continues to be the basis of his work. Like most others, he has no idea how long this phase will continue, but based on that analysis, he is starting to hone in. If he has learned nothing else, it’s this...
I think around the bull market it’s really important to understand the psychology of how it develops. When something big and visible has been going on for a long time – it becomes obvious to everyone. Everyone comes to believe that it’s the road to riches. In 1999 I’d go on lunch dates with 20-something’s and all they wanted to talk about is their technology mutual fund. My college roommate called me to tell me he was making lots of money on a Mexican internet stock called Cisco. Now almost all the young people I talk to are starting or being involved in AI startups.
He adds...
Between governments racing to get the lead in AI, all these people getting into AI startups, almost every company at the minimum commenting on AI and the kinds of price action we are seeing in fairly large stocks – I think we must be getting close to the start of that ending stage. My biggest struggle is - what’s the correct proxy to follow? Is it NVDA? Is it QQQ? Some basket of PLTR and those types of names?
This is all part of trying to solve the trillion dollar question because while everybody can hash over what has and hasn’t changed, one thing that hasn’t and never will is this: The markets can be very humbling... to everybody.
▶Speaking of which, I received this email Friday night from someone who had seen a preview of my “CoreWeave” write-up, without seeing the full report...
With all due respect, you’ve been dead wrong absolutely dead WRONG for months. It’s been painful. Nobody gets it right all the time but don’t have such great convictions. Think about how you were so down on MP. Oh boy.
Ah, this is about MP Materials $MP, whose stock I had red-flagged, and then went on to be saved after the government took a major stake. He must’ve missed my follow ups, including the one where I wrote...
I always say, if an investment thesis hinges on the outcome of courts, Congress or a regulatory body – good luck! Maybe going forward I’ll add, ‘The chances of the U.S. government taking a stake.’
All of which is a long-winded way of saying: It’s not just the markets that can be humbling. So can hubris. We’re in that phase. Time to rev up the ole’ Hostile React-O-Meter™.
As always, I’m working on new Red Flags. Until then, have a great week.