If you are a $CRWV investor, you may find this article helpful to understand how CoreWeave's business model works. Leave a comment if you have questions or suggestions. https://t.co/QY0K5HHmeh
$CRWV CoreWeave has earnings on Tuesday after the market closes.
While I'm not (yet) making concrete predictions (though I did drop a few numbers in the earnings preview video coming out tomorrow morning), here's what I'm expecting.
🟢 REVENUE
Analysts are expecting $2.555b in revenue, and this continues the 100% Y/Y growth trend. They will likely beat this by a small amount due to the timing of new projects and capacity coming online, though it's a tight range by analysts. Maybe a 5% beat. Not the most important # in these earnings.
Their ARR goal by the end of 2027 is $30b, and they're ahead of schedule as the backlog numbers they posted indicated around $36b of expected deliverable supply through Q1 2028.
🔴 EPS
EPS is a wider range than you'd find with most companies, varying from -.42 to -1.82, with -1.21 as the average. This relates to CapEx, and due to the reads of $GLXY and $APLD and the recent underwriting of the bond raise, it looks like this is going to miss. I'm guessing we're at a -1.4 or so, barring some one-off benefit that helps their EPS.
Historically, they've missed on EPS 4 of their 5 quarters, so I'm not sure this will rattle the market much. It's more important right now to the market, but the longer-term outlook is what's most important here.
🟢BACKLOG
Backlog is set to explode. Last quarter, it was $99.4b. Bloomberg estimated $105b back in June. Cantor-Fitzgerald had an excellent note touching the backlog based on the prior bond offering memorandum. They expect somewhere between $125-$131b, a sizable beat, and I think that's reasonable.
🔴 DEBT / CapEx
To secure the backlog, debt and CapEx is necessary. This will probably be a bit ahead of estimates, due to the robust backlog and the ramping business. The CapEx estimates are a tremendous range, from -$14.2b to -$6.3b, average -8.3b.
It's really tough to say on this what the number will fall at, as it appears to be somewhat lumpy, though directionally accelerating. In 3 of the 5 quarters CoreWeave has reported, analysts overestimated CapEx by an average of 43%. In the prior quarter, CoreWeave missed to the downside by nearly 30%.
STOCK REACTION
CoreWeave remains undervalued despite the recent surge. I think much of the reaction will depend on where that ARR is sitting, the backlog, and of course the debt. They may also announce a new contract (though '26 is essentially sold out) which might also add some momentum.
I'm guessing we see CoreWeave move up some, but as many are irrationally afraid of the debt (which is dwarfed by the backlog), it may not be as much as would be deserved.
AI has definitely excelled exponentially in the last two years and I can comfortably say that because I haven’t hand coded things for last year and a half, except for the first 3 months of 2026 when I was grinding leetcode.
You could tout and say you have built X, Y, and Z but these interviews test your real knowledge and test if you understand the system you built deeply, the tradeoffs you made, and the lessons you learned. There’s no way you can comprehend any of that unless you are the master of your craft.
@ChrisCamillo I spent 6 hours just yesterday trying to understand the business model and debt structures of $CRWV. Read all the 10-K filings from SEC and company resources and articles. I am a lot more informed investor today than I was yesterday.
If you are a $CRWV investor, you may find this article helpful to understand how CoreWeave's business model works. Leave a comment if you have questions or suggestions. https://t.co/QY0K5HHmeh
Will there be a time where we have excessively over built the data center capacity adding to the so called bubble? Possibly. But market will get a whiff of that long time before it actually happens because if how information proliferates today. This is not the same market where only elitists from top firrns had access to information like in the year 2000s.
Short sellers like Dr. Burry is actually keeping the market from bubble, so I appreciate his participation.
@michaeljburry I could refute and say you are still wearing your old glasses from 2008 to look at the world in 2026 but like everyone else, I could end up in Michael Burry Blocked List. Hence, I will shut up.
@gdb I don’t know about the benchmarks but my coding harness that I built on top of Codex sometimes accidentally switches to Luna and it takes me just couple prompts and answers to realize that the answers aren’t the same as Sol. What am I doing wrong?