re: Nebius - NBIS - AI Infrastructure Company

So my last comment on this cluster of a earnings release and call, and it may seem like I am being harsh and it is because I am. They trumpet themselves as best in breed management and get a premium for it - then they do this. The presentation decisions they made and lack of transparency was just terrible. Particularly the decision to stop projecting ARR - let me walk through this.

Q3 revenue - $958 million, of which $300 million is one time - so running revenue $658 million. So truly compared to consensus (which did not assume any one time revenue) this is a big miss due to delays they hid and suggested did not exist.

Q4 ARR they downgrade to $3-$3.4 billion. Remember at Q2 26 they started ARR for THAT quarter was $3 billion. So really not good.

OK fine - it is what it is - they have 2027 to paint the real demand and growth picture. And what do they decide to do:

1) Disclose CAPEX of $40-45 Billion
2) Say they will no longer guide to ARR
3) Say total revenue will be $12-13 billion.

What the hell is the market supposed to think of that? First ARR is critical to substantiate why we are spending $45 billion. An inexperienced investor will see a $3 billion ARR coming in and then see only $9 to $10 billion in incremental revenue on $45 billion in CAPEX. That is what ARR helps to explain (ie that such a comparison is not apples to oranges). But that is gone now - the market is on their own.

And lets unpack $12-13 billion in revenue. $3 billion ARR coming in and promises of over 1 GW of new capacity coming online. New capacity is trending to get low 20's on LT deals and 40-50 on short term. Given they say $4 billion in ARR in Feb (the delayed 12/31 ARR) - this suggests the extra GW will contribute $8 to $9 billion? That must suggest this 1 GW is coming on quite late in the year (as a midpoint convention would suggest dramatically more of that) plus any asset light as there is zero of that in it. ARR projection would make that clear - it would say sure revenue will be $12 to $13 billion but our exit velocity will be closer to $30 billion. But no guidance.

Margins drop in Q3. Company projects flat margin in Q4 and no guidance in 27. Analyst asks question on it and management fumbles and says one off revenue causes static so ignore Q3 (which means Q3 was actually really worse if it was not for the one off revenue) and fumbles going forward instead of blasting that the 40% margin is because NBIS does not own their capacity, but 75% of the new capacity they will and that will skyrocker margins. She mumbled something of the sort at the end but the horse was out of the barn.

Inexcusable presentation. Inexcusable action by Data One misleading everyone on Vineland in such a public fashion. Really a bad day.

Bullish AF though because I see through the noise. Iren's historically horrific earning calls (last one was good) have trained me for this crap - just never expected it from the A team.

quote:

---

Bullish AF though because I see through the noise.

---

I bought this and plan to hold until 2029, at least. The thing I'm looking for is the AI demand. I will consider selling if the demand starts dropping, but that hasn't happened, yet. We're still in the "Can't build fast enough" stage.

I am also looking for that stage where they can't build fast enough, BUT the majority of their customers are not AI companies, but regular companies looking to build up their AI processes.

My daughter works for a medium sized construction company in Portland, OR. She is their technical lead, and is frustrated because she is getting pushback from folks who are old school and want to keep doing things the way they always have. There is some inertia that needs to be overcome to get companies to change the way they do things, and that is kind of where we are now. The early adapters are going to get a competitive advantage and everyone else will wake up and look to companies like Nebius to help them compete.

quote:

---

An inexperienced investor will see a $3 billion ARR coming in and then see only $9 to $10 billion in incremental revenue on $45 billion in CAPEX

---

lol…You mean smart investors that don’t beiieve the magic beans these people are claiming will exist in 2030.

If you are spending 45 billion, to make 10 billion, that’s not a sustainable business. Ed Zitron has been making this point forever on the neo Clouds as well as OpenAI/Anthropic.

Especially for a stock like NBIS that trades at excessive valuation and has still 10x’ed in just 2 years. Where else do you think the stock will go from here? You could make the argument two years ago to buy because the argument of “hyper growth will allow us to scale”. Well it’s two years later and there is no scaling going on. It’s still a bankrupt business model.

They stole my money that I invested in earlier this week!!

quote:

---

And thereby proving my point of how this gets misinterpreted and twisted when not explained

---

I’ll explain it with real numbers. We can just look at 2 year intervals to see if there’s any scaling happening.

NBIS in 2024/2025, from their own Annual report, spent a total of 4.888 Billion in CapEx.

NBIS 2026 Full Year revenue guidance is 3.0 Billion.

2025/2026: Cap Ex- 31 Billion
2027 Rev: 12.5 Billion

4.888/3.0=1.629
31/12.5=2.48

They are doing the opposite of scaling. Spending more to bring in less. And that doesn’t even factor in this is just revenue. It doesn’t factor in the operating costs themselves. If you invested in 2024, you would not have expected them to go in this direction as they expanded. Certsinly not in an interest rate high environment. CapEx isn’t free

This isn’t a real business. It’s why the short shares are 21% of the float. I expect that number to increase after this report.

Are there no operating expenses? Why do you guys keep quoting revenue? Since when does revenue = profit?

Loading Twitter/X Embed...
If tweet fails to load, click here.

Earth shattering analysis.

quote:

---

I understand you don't understand long-term construction projects

---

Oh I understand them plenty. There’s a reason they don’t escalate in value 10x within 2 years of construction. It’s nonsensical.

Many more shoes to drop, it’s never just one release

They are going to do a huge dilutive equity issuance in the near term and the stock will drop 10-20%

And that’s the bull case, because at least there’s enough compute demand there to justify the capex