Nebius just reported 25Q4 earning. Although revenue missed consensus, Nebius addressed almost all my concerns with the company in this earning call and I think the growth path of the company is now clearer than ever.
Before the earning call, here were my primary concerns:
- The company guided $7B ~ $9B ARR by end of 2026, but how much CapEx would they need to achieve this goal? What will be the funding source? Will there be any significant dilution or CoreWeave level high rate debts?
- It appeared that DataOne (Nebius’s colocation partner) had 2 months delay on their initial NJ data center project. And many people flagged that time to power was a big bottleneck for this industry. Given all these conditions, will Nebius succeed in delivering the projects?
- Is renting out GPU really a profitable business? What will be the margin? What will be the unit economics in the hyperscaler deals, especially compared to peers like $IREN? Will Nebius’s profitability be limited due to colocation (VS $IREN who builds data centers end to end)?
- As a respectable member on this board pointed out, bare metal deals with hyperscaler might dominate Nebius’s short-term income and forced it to become an IaaS company, which was against the company’s initial strategy. Will the software stack and cloud approach really add any value?
Here are the answer I found to each question from the earning call:
Regarding CapEx, funding, debts and dilution:
Nebius provided guidance of $16B - $20B on CapEx but claimed that approximately more than 60% of this will be covered by cash flow. So there will be only $6.4B to $8B that needs funding. Because Nebius does not currently have any corporate debts or assets based financing yet, it appears that these would be their preferred choices rather than more dilution. And Nebius clarified that they had not used any of the ATM offering they authorized last time yet, so this would be another flexible source of funding depending on the share price performance.
These information provided a lot of clarity. Although we still don’t know the actual term of the debt offering yet, this at least sets a clear expectation for investors to model their business.
How are we going to finance the CapEx? So obviously, we will first finance it from our cash flows. We have our cash on hand. We have cash that is generated from our core business. But most importantly, I would say, we have significant amount of cash that we received and will continue to receive in 2026 from the favorable terms of our long-term contracts. We are talking about a significant amount of cash.This cash flows will actually finance the majority, actually around 60%, maybe even more of all of our CapEx needs in 2026.
So how are we going to finance the remaining amount, the rest? So as we all know, we have a very healthy balance sheet. By the way, not by coincidence because we are working very hard to have a very prudent, disciplined, healthy balance sheet. As of today, we don’t have any corporate level debt. We don’t have any asset-backed financing, even though we have multibillion dollar revenues from long-term contracts. We don’t have any bank revolver. And we didn’t have it to date by choice.
But moving into 2026 and as we all know, an optimal capital structure in our business should include also debt. This should obviously be changed. And we plan during 2026 to use some of the tools that I mentioned in order to move towards a more optimal capital structure. So having said all of that from the financing point of view, we believe that the $16 billion to $20 billion CapEx makes a ton of sense, and we will be able to finance it while keeping a very healthy disciplined balance sheet.
As you all know, we launched our ATM program last quarter. As of today, we didn’t use it at all. Actually, we don’t have any concrete plans to use it also in the near future. However, it’s another tool in our toolbox that will enable us to use it when it will make the most sense for our business as well as our shareholders. And this is another tool that will enable us to keep a prudent balanced disciplined balance sheet.
Regarding execution and bottlenecks:
Nebius’s ending ARR of 2025 came at $1.2B VS $0.9B - $1.1B guided. And Nebius emphasized that both the Meta deal and Microsoft deal are on track, with the Microsoft deal having “some buffer time“. The full capacity of the Meta deal has been delivered since February while the remaining tranches of the Microsoft deal is expected to be delivered in 2027. (DataOne’s CEO once said that they would deliver by November 2026, so this might be the buffer time Nebius was talking about. ) At the same time, Nebius is bring online 9 new data center sites across geographies and increased the guidance of contracted capacity to >3.0GW by end of 2026. Nebius said that their Q1 capacity is already sold out and the sales pipeline is on track to exceed $4B in Q1.
These are all really good examples of solid execution.
We are expanding our global footprint with nine new sites across seven locations in the US (Missouri, Alabama, Oklahoma and Minnesota) and EMEA (France, Israel, UK), bringing our global total number of sites to 16 and broadening our geographical diversification.
In terms of deal trends, they’re all moving in the right direction. Deal terms are getting longer and average deal sizes are increasing. In Q4, we saw nearly twice as many transactions completed over 12 months in duration over what we succeeded with in Q3, while average selling prices increased by more than 50%. As an example, we are sold out of Hoppers and those that are coming up for renewal often off of short reserve agreements are getting renewed at 12 months or longer, while we’re actually seeing pricing nudging up. Lastly, we’re focusing on customers with premium workloads and use cases, which is resulting in superior terms, including increasing those who are willing to prepay for securing future capacity.
As of early February, Meta’s capacity was fully deployed, and we are now fully servicing their contract. After delivering the first tranche of our Microsoft commitment on time, we expect to continue to deliver the remaining tranches throughout the year with the majority expected in the second half. We expect Microsoft to begin contributing to revenue at the full annual run rate in 2027 once we have deployed all of these tranches
About profitability and unit economics:
Nebius guided 40% Adjusted EBITDA margin for 2026, which looks a very good number considering its early stage. Nebius reiterated their mid-term guidance of 20-30% EBIT margin.
In terms of unity economics, since $IREN disclosed all details about their Microsoft deal, I always wanted to compare it with Nebius’s Microsoft deal. $IREN’s deal is 200MW IT load, $9.7B in 5 years, while Nebius’s deal is 300MW IT load, $17B in 5 years. So a napkin math can easily tell that Nebius’s deal generates ~17% more revenue per MW than $IREN’s. Because $IREN also managed to negotiate a $2B prepayment, it was still impossible to compare them apple-to-apple. But in the financial statements of Nebius’s earning report, we can find that there is ~$1.5B deferred revenue added in Q4 (as shown in the following screenshot). Speculatively, I think this should be from prepayment of either the Meta deal or the Microsoft deal or both.
Considering this additional information, I tend to believe that Nebius’s deal has better unit economics than $IREN’s, if we exclude the colocation cost of Nebius’s data center. Nebius also said that they would prefer owning end-2-end data centers moving forward while still maintaining a portion of colocation to expedite execution.
Generally, we are focused on bringing most of our largest projects as self-developed projects. There are 3 main reasons. First, we get much better ownership with this. We have much more control over the execution of the project because we have expertise and experience of building our own sites.
We also can achieve greater efficiency at scale because generally, we tailor the design specifically for what we need and for our technical requirements. But this takes time and while we use leases and partnerships to fill the gaps before we are at a full speed with our own one. So just to sum up, the preference is to develop the infrastructure ourselves. But again, we still need the partners and leases occasionally to support the growth.
About the value of software and end to end cloud:
Nebius explicitly said that they do not want all deals to be hyper scaler. Their vision is to use software offerings to attract AI native startups / enterprise, and to focus on building vertical solutions (e.g. healthcare).
Although AI native startups all start small, some of them can grow really big and really fast because of the disruptive technology. And Nebius has already observed some trends like this.
We are, as you know, much more focused on our AI cloud business. And in AI cloud, we have these 2 major sectors, AI start-ups and enterprise and look at AI native customers, what happened to them in 2025. Some start-ups disappeared, but some of them are becoming real companies, real enterprises of the future. And they started getting traction. Their products became more and more – they’re more and more used by their customers, and they’re scaling quickly, scaling with real revenues, real demand.
And what – we see such customers, such companies who used to order hundreds of GPUs, thousands of GPUs, now they’re ordering tens of thousands of GPUs. And this is actually is a magnitude of what we saw in the largest consumers last year. Frontier models were ordering tens of thousands of GPUs just a year, 2 years ago. Now it’s yesterday’s start-ups at this level.
Again, this is real business. These are real customers of them paying real revenues. So AI traction is visible. I can – there’s a lot of customers from the sector, starting with Cuckoo, Cursor, Rodo, Higgsfield, Photoroom, Genesis Molecula, different sectors, and they all have their first traction and they’re becoming – as we speak, becoming real, real companies. Actually, it’s the future enterprises.
And on the other hand, there is enterprise clients who involve more – who actually switching most of their everyday business process to AI and generate new profits through that AI implementations. We see the growing number of such customers, growing contracts from each of such a customer, the number of GPUs is growing, the duration of the contracts is growing. As I said, the average duration for new customers grew 50% last year.
When being asked whether the 2026 ARR guidance will be dependent on any additional hyperscaler level deals, Nebius said that they would like to balance their large deals and AI cloud deals. I believe this is the right direction and the diversification of revenue can provide a lot of safety to Nebius’s future sustainability.
Our 2026 ARR target is not dependent on any new mega deals. As we bring on our planned additional capacity, combined with the already strong pipeline, and our go-to-market plans, we are very confident in our ability to deliver our '26 ARR target.
Our continued success with AI natives and the early progress we’ve seen with ISVs and enterprises set the foundation for capturing the market this year. Based on the traction we are experiencing and our extensive research on our total market opportunity, we are leaning into the verticals we’ve already laid out, health care, life sciences, media and entertainment, physical AI and retail, which give us ample runway to capture share and grow our business. While we are happy to service large strategic customers like hyperscalers, we will remain opportunistic with such large deals as we look to balance the opportunity with the long-term positioning we plan to achieve with our AI cloud.
Nebius intends to be the AI cloud partner of choice for both cutting-edge startups and larger enterprises. We help solve the biggest problems our customers face by delivering reliable capacity at scale, providing a comprehensive, full-stack AI cloud platform, and consistently expanding our offering with new products and services to meet customer needs.
Our ability to deliver AI compute and cloud services quickly and efficiently has won the trust of the leading AI-native start-ups.
Our enterprise-ready solutions, which meet the unique security and compliance requirements of complex organizations, are increasingly being selected by larger enterprises.
Our focus on high-potential industry verticals is gaining traction from our ability to optimally configure leading-edge compute performance, networking, and storage, all in a fast and convenient out-of-the-box package based on specific customer needs.
We have already gained strong traction in healthcare and life sciences, where training, storage, and HIPAA compliance are paramount. Innovative early-stage businesses like Genesis Molecular AI, is building their AI solutions on Nebius.
Companies at the cutting edge of physical AI and robotics are leveraging Nebius as they work to bridge the gap between the world around us and AI. These customers need training compute power to create simulated environments, and a significant amount of storage. Nebius can flex its offering to resize and reconfigure our solutions to meet these unique needs.
We are working with World Labs, which is building the next frontier of generative AI where models can understand and interact with the world, and we are helping a stealth frontier research lab building an advanced robotics foundation model.
Customers in media and entertainment need high-performance compute and large-scale storage for model creation and inference to support media creation from still images through to full frame video. Recent examples include Higgsfield AI, which leverages Nebius to deliver stylized video content with true cinematic control, while Photoroom automates AI product visual creation for businesses.
To capture the vast opportunity ahead, we are building a worldclass sales and GTM team. In 2026, we will:
- Focus on key enterprise industry verticals, including healthcare and life sciences, physical AI, and media and entertainment, as mentioned above. We are also growing our footprint across the retail/e-commerce, and financial services verticals. These end-markets are seeing early traction in AI adoption and we believe they will drive substantial demand for our AI cloud, creating significant long-term market potential.
- Expand our GTM leadership team and organization to better capture the immense global demand in front of us.
- Enhance our engagement with the developer community.
Cheers,
Luffy