CoreWeave IPO, Wolfspeed Reorg, and Alphawave Bid

This is a shorter piece talking about some of the cross currents in markets right now.

It’s been a while since I’ve had some rapid-fire takes. Reminder: Be gentle with me. I’m recovering from surgery and a bit off my game. I know this is turning into a macroeconomic newsletter these days (jokes!), but I do want to talk about some big moves in the space.

First, let’s talk about Coreweave. As a friendly reminder, it was on Core Research on SemiAnalysis before it was here.

CoreWeave IPO - Hated but a Lesson in Crowding

CoreWeave is… contentious. I have not heard more pessimism in a new offering in a long time. The bankers were skittish, and most investors wanted to short it out of the gate. Despite the recent ClusterMax Article by SemiAnalysis saying they are the best, investors disliked the terminal value issues, the lack of clear new incremental customers, and the heavy adjustments. The most cynical take is that these are like airline leasing businesses with worse terminal lives.

But despite all this, SemiAnalysis was broadly positive. I am, too, but probably not for the exact same reasons. They are the best at what they do, and while renting GPUs might end up as a leasing business, today, they have a legitimate competitive advantage that you cannot buy. Hyperscalers have worse GPU experiences, and I earnestly believe the emperors are wearing no clothes. No one had true GPU at the 10k+ level until recently; CoreWeave has one huge advantage on its side: focus.

I understand investors' concerns, but what’s worse is investors have been so obsessed with philosophical aspects of the business model that they are missing the accurate big picture: the setup to short sucks. It’s crowded, the borrowing is insane (150%+ annually), and the underwriters can buy into the offering to move the price. When the consensus is so heavily in one direction, often it pays to do the opposite.

What’s more it’s almost impossible to predict end demand. In their roadshow, they talked about how large customers tend to expand. While Microsoft is likely not expanding, the largest GPU buyer has just become their customer. I would expect new contracts to drop from the sky on investors (who are heavily short) and somehow hurt this dynamic even more. At specific prices I would expect a secondary almost immediately. This is to say, this will be a battleground stock, and the capital stack re-rate will help CoreWeave in the long run.

Anyway, that’s my take on CoreWeave’s IPO. I sort of am sympathetic to the concerns in the long run, but it’s no surprise they are IPO’ing at this time when they are about to ramp GB200s, have the largest first allocations, and just landed a deal with OAI. What’s more, this is extremely low float and super hated. This screams like every other semiconductor IPO in the last two years. ARM and ALAB had similar setups into earnings, and the first earnings had massive beats that murdered short sellers.

I would not be surprised if history repeats itself almost precisely. Anyway, let's move on to Alphawave’s bid and Wolfspeed’s reorganization behind the paywall.