Friday File: Portfolio Construction in a Time of Greed and Fear

Holy cow, Cerebras. What a wild year.

I wrote quite a bit about Cerebras Systems (CBRS) and its wafer-scale engine (WSE) chips last year, when several newsletter guys, beginning with Jeff Brown, started pitching the stock as a buy going into the then-expected IPO… and, sadly, I found the potentially really intriguing, and I spent a few weeks negotiating a purchase of some CBRS shares on Hiive as their IPO was delayed, and had a deal lined up to buy at somewhere in the $35-40 neighborhood, but I ended up deciding not to buy.

So that sure looks like a missed opportunity — now we’ve had another year of wild AI enthusiasm, with dramatically more optimism than there was even twelve months ago, and Cerebras went public at $185 this week, and traded up to $380 or so for a few minutes on its first day. So I passed up on a 10X return in a year, more or less.

Dangit.

The valuation is completely nutty by most rational measures, of course, Cerebras is now a ~$70 billion company, and the strongest argument is that they have a backlog of over $25 billion… which sounds better than saying they have trailing revenues of ~$500 million, and are valued at 140X 2025 revenues. But expectations will probably shift quickly this year as their profile rises, and as analysts start publishing estimates in a month or two (and as insiders are able to start selling, usually in about six months). There’s a lot of crazy-sounding valuation going around, and it doesn’t mean any of those richly valued stocks will stop looking nutty anytime soon. This is all about the future being almost unimaginably bright, and maybe it will be, certainly nobody thought NVIDIA was going to become a $5 trillion company ten or fifteen years ago (or almost $6 trillion now)… just make sure you’re not betting all of your portfolio on just “unimaginably bright future” stories.

And I will note that NVIDIA was a $70 billion company as recently as 2017, roughly the size of Cerebras, only eight or nine years ago — so the optimistic take might be that such a company can explode into something far larger in a pretty brief period of time, especially if it happens to be the best positioned company for the largest capital investment cycle in the history of the world, as was NVIDIA’s …