If the price is > US$100, I might take a small speculative position, but not full allocation.
1. Company Overview
1.1 What does the company do?
Nebius is a technology company building vertically integrated infrastructure for the AI industry: cloud platforms, large-scale GPU clusters, data centres, and related tools (for training/inference, etc.). It also operates some subsidiaries/brands such as Avride (autonomous driving), TripleTen (edtech / reskilling), and has equity stakes in Toloka and Clickhouse. (Nebius)
1.2 Sector/Sub-sector
Sector: Technology / AI infrastructure / Cloud computing. Sub-sectors: AI cloud services, data centres, hardware + software tools for AI workloads. Edtech and autonomous vehicle tech are adjunct models. (Nebius)
1.3 Total Addressable Market (TAM)
Data is less precise in what I found. But since AI infrastructure is exploding, the TAM is very large: includes AI training/inference, cloud services, GPU compute, data centres, possibly edge compute, etc. Nebius is aiming at “AI builders from startups to research institutes to enterprises” globally. (Nebius)
1.4 Main Competitors & Market Share
Competitors include big cloud providers (AWS, Microsoft Azure, Google Cloud), AI infrastructure startups, and hardware-focused compute providers. Also other specialized AI infrastructure firms. Nebius is still small relative to giants, so its market share in global cloud compute / AI infrastructure is minor. I didn’t find good data for exact market share.
1.5 Dependence on Single Product/Customer
Doesn’t appear to be heavily dependent on a single customer. Its offerings are broad. It is building several verticals. But risk exists: high capex, hardware cost, performance, reliability, and competition.
2. Quick Screen
2.1 Debt: Current debt-to-equity ratio
~ 26 % debt to equity. (Simply Wall St)
Total debt ~$986.2M, equity ~$3.78B. (Simply Wall St)
2.2 Registered address
Headquarters in Amsterdam, Netherlands. (Nebius)
2.3 Scandal/Fraud: Records involving the company
I didn’t find any major fraud scandals. There was the restructuring: Nebius used to be Yandex N.V., and due to sanctions, it was split and sold off Russian parts; now focusing on non-Russia AI businesses. There’s potential geopolitical risk. (Wikipedia)
2.4 Scandal/Fraud (Promoters)
No obvious recent promoter fraud that I saw. But the change from Yandex to Nebius due to sanctions means there is history of regulatory stress. The promoters had to restructure. That may impact trust / regulatory risk. (Wikipedia)
2.5 Debtor Days
I did not find data on debtor days (receivables turnover etc.).
2.6 Market Cap
Approximately US$24.8-$26.6 billion. (Companies Market Cap) This is large-cap by most definitions.
2.7 Cash Reserves
Cash on hand ~$1.68B. (Simply Wall St)
2.8 Free Cash Flow (FCF): Positive & Growing?
The analysis suggests its operating cash flow is negative. So FCF is likely negative (or borderline), and not yet positive/growing in a stable way. (Simply Wall St)
2.9 Foreign Exchange Exposure
Being headquartered in Netherlands, operating globally, using hardware / GPUs, etc., it has exposure to USD, EUR, other currencies. Also supply chains in different locations. But I didn’t find a specific quantification of forex risk.
3. Shareholder Alignment
3.1 Past allegations vs company/promoters
None major that I found, apart from the geopolitical/regulatory issues associated with its history as Yandex.
3.2 Promoters: Who are they, and their track record
Founded by Arkady Volozh, who was also involved with Yandex. He has experience building large tech companies. The shift in structure (splitting from Yandex, dealing with sanctions, etc.) shows ability to adapt but also introduces risk. (Wikipedia)
3.3 Promoter Ownership
I did not find a clear figure for how much promoters still own, or recent trend.
3.4 Pledging
No data found on share pledging by promoters.
3.5 Openness: Management Transparency
They have public filings (SEC), they disclosed the raising of capital, they do investor communications. Some transparency. But risk remains around what parts of the business are more or less mature or loss-making.
3.6 IPO / Use of IPO</Red Herring
Not exactly relevant since NBIS was previously Yandex N.V., which had IPO. The recent capital raises (Class A shares, convertible notes) have been used to raise funds for growth / constructing infrastructure. (TradingView)
3.7 Institutional Investors
Not much detailed data in what I saw about who holds what % (FIIs, DIIs etc.).
3.8 Insider Trading
No obvious reports of insider malfeasance.
3.9 Dividend Policy
No dividend – it’s growth status, and likely reinvesting capital, so no regular dividends. (INDmoney)
4. Performance
4.1 Current share price
About US$106.58 (recently around there) per share. (INDmoney)
4.2 PE Ratio
Very high: ~ 110-140× depending on which metric. (Simply Wall St)
4.3 Sales / Revenue
Recent trailing-twelve-months (TTM) revenue ~ US$249.3 million. (Simply Wall St)
4.4 Profit (PAT)
Not very strong; earnings are positive recently but very small relative to market cap; profits are not consistently strong. Also history of operating losses. (Simply Wall St)
4.5 EPS trend
EPS is positive recently (~0.76 per share TTM) but since there are recent losses historically, it may be volatile. (Simply Wall St)
4.6 Margin Stability
Margins (gross margin ~53.7%) are decent in the revenue → cost of goods sold sense. But net margins vary, and likely negative in some periods. Stability is questionable. (Simply Wall St)
4.7 Segment Performance
Main revenue comes from its AI infrastructure/cloud compute. The other segments (edtech, autonomous driving) are smaller / growing but not yet materially contributing large revenue (from what I saw).
5. Efficiency
5.1 Operating Profit Margin (OPM)
Data is mixed; they’re still investing heavily. It’s not clear operating profits are positive or stable.
5.2 ROCE (Return on Capital Employed)
No specific number found. Given high capital expenditure (capex) and depreciation, returns may be low now; expected to improve if the infrastructure starts being utilized fully.
5.3 ROE (Return on Equity)
Some positive recent earnings, but given equity base large and earnings small, ROE is modest.
5.4 ROCE vs ROE
Insufficient data to conclude, but likely ROCE < ROE or both low, given early-growth investment phase.
5.5 Debtor Days
No data.
5.6 Asset Turnover Ratio
Not found.
5.7 Working Capital Cycle
No data.
5.8 Inventory Days
Not applicable / minimal inventory for cloud infra business.
6. Financial Risk
6.1 Debt: Long-term or short-term concerns
Total debt is ~$1.2B; debt to equity ~26%. Also, cash reserves ~$1.68B, which gives cushion. (Simply Wall St) The risk is whether the growth & utilization can justify the high capex & maintenance of the infrastructure.
6.2 Pre-IPO Debt
Not applicable exactly; major recent raise via share offerings and convertibles, rather than heavy leveraged debt.
6.3 Future Debt Plans
They just raised ~$4.3B via offerings (public + convertible notes) to fund expansion. That’s a lot of capital, which dilutes or adds financial obligations. (TradingView)
6.4 Credit Rating
I did not find a credit rating (Moody’s, S&P etc.) for Nebius.
6.5 Contingent Liabilities
No obvious large off-balance sheet liabilities found in the sources I checked.
6.6 Legal/Regulatory Risks
Yes: history with Yandex / sanctions. Geopolitical risk especially in relation to Russia, regulation of AI infrastructure, export controls (especially GPUs, etc.), possible supply chain constraints.
7. Volume & Liquidity
7.1 Today’s volume
Trading volume is fairly active for a stock this size. I saw tens of millions of shares volume in recent days. (INDmoney)
7.2 Shareholder base
Not detailed. But with public offerings and convertible note issues, institutional involvement appears strong.
7.3 Promoter Pledge Trend
No data about pledging.
7.4 Institutional Exit
No strong evidence yet of institutions exiting en masse; some may be wary of valuation.
7.5 Volatility
High volatility. Stock has moved from ~$14 (52-week low) to ~$100+, so that’s huge gains in a year. (INDmoney)
8. Valuations
8.1 PE Ratio: Current & justified by growth?
PE is extremely high (~110-140×). To justify that, Nebius needs massive growth in revenues and profit (high topline growth + margin expansion). Some of that is priced in. (Simply Wall St)
8.2 PAT Growth vs PE: GARP logic
If earnings were growing fast (they are growing, but from low base), maybe some GARP case. But growth has to be sustained, and margin expansion significant.
8.3 Market Cap to Sales
Market cap ~$24-26B vs revenue ~$249M → P/S ~ 100×. Actually in some sources P/S ~106.7×. Very high. (Simply Wall St)
8.4 Price to Book
Not found clearly; but given large equity base, probably high also.
8.5 Dividend Yield
Zero. No dividends.
8.6 Market Cap Growth vs Fundamentals
Stock has rallied massively, likely more in expectation of future growth than current fundamentals. So fundamentals are lagging stock price.
8.7 PEG Ratio
Not enough reliable data / consensus growth rate to compute. Probably PEG is high (since growth expectations large).
8.8 EV/EBITDA
Could be very high or not meaningful if EBITDA is negative or small.
8.9 Historical Valuation vs 5-year average
Since company became NBIS fairly recently (after Yandex split etc.), historical data is limited; but stock has soared from very low base; valuation multiples are much higher than typical tech / cloud / AI infra peers.
9. Growth Components
9.1 Market Cap: Growing consistently?
Yes, large growth over past year. (Companies Market Cap)
9.2 Sales Growth: Y-o-Y
Strong growth but from small base; the challenge is scaling.
9.3 Profit Growth: Sustainable with margins?
Profit is small; sustainability depends on cost of capital, infrastructure utilization, pricing of AI cloud, competition, supply chain, energy costs etc.
9.4 OPM Growth: Improving margins over time?
Potentially, as scale improves, but currently margins (especially net or operating) are volatile and suffering from investment costs.
9.5 Capex: Investing in future growth?
Yes, big capex / capital raising (convertible notes, etc.), building data centres and clusters. (TradingView)
9.6 Moats: Lasting advantages
Potential moats: integrated stack, ability to design hardware + software optimized for AI, geographic spread of data infrastructure, relationships with developers. But big risk: competitors with deeper pockets (Microsoft, Nvidia, Amazon) can out-invest or under-price.
9.7 Customers: Diversified & recurring?
Likely diversified; developers, AI labs, enterprises. Recurring usage possible. But contract detail not found.
9.8 Employee Base: Growing
I saw ~1,371 employees per one source. (Google) Likely growing given expansion.
9.9 R&D Investment
Yes: designing own hardware/software optimizations. But exact R&D spend not clear in what I saw.
9.10 Expansion Plans
Data centres under construction, expansion of GPU cloud platform, expanding global footprint. (Nebius)
9.11 ESG / Green Transition
No strong info found about sustainability goals, energy efficiency, carbon usage, etc. Might be material given data centres are energy intensive.
9.12 Management Succession
Arkady Volozh is CEO; key leadership seems stable; but company is young in its NBIS form, so succession planning less visible.
Summary: Strengths & Risks
Strengths
- Operating in a very hot, high-growth market (AI infrastructure).
- Strong market capitalization, ability to raise capital.
- Cash reserves more than debt; debt is not overwhelming at this stage.
- Integrated approach (hardware, software, cloud platform) could lead to competitive advantage.
Risks
- Valuations are very rich; huge expectations baked in.
- Free cash flows negative; profitability uncertain and may be volatile.
- Intense competition from giants with more resources.
- Regulatory/geopolitical risk (export controls, sanctions, supply chain, etc.).
- Capital‐intensive business; need high utilisation of infrastructure to make money.
Purchase Price / Valuation Suggestion
Using a Buffett/Munger style, I think a margin of safety is essential here. Given the current valuation multiples and risks, you should base your buying price on conservative projections.
Here’s how I’d think:
- Assume revenue grows at, say, 3× to 5× over next 3-5 years (just hypothesis), and margins improving but not explosive.
- Use a discount rate reflecting tech risk + execution risk (let’s assume ~10-15% given AI infra, etc.).
- Look for a price where P/E becomes more justified; maybe when P/E compresses to ~30-50× (if growth is strong) – current ~110× is high.
So my recommendation: If I were buying, I would only be comfortable entering a meaningful position if the stock price drops to a level corresponding to valuation of, say, US$60–80 per share (this gives buffer), assuming growth proceeds roughly as hoped.
If the price is > US$100, I might take a small speculative position, but not full allocation.
Frameworks / Thinking Tools to Apply
Here are some more advanced frameworks (in Buffett / Munger spirit) to further analyze this:
- “Earnings Power Value” plus Growth Discount
Estimate what Nebius can earn once it reaches stable capacity utilisation (say 3-5 years out), then discount that back to present using a conservative rate. Compare that with current price. - “Optionality” value
Nebius is partly an option on AI compute going vertical. That optionality has value but should not dominate your valuation unless very likely. Separate the “core predictable business” from the “moonshots” (edtech, autonomous driving etc.) - Capex Sustainability & Utilization Curve
The financial return depends heavily on how quickly the data centres / GPU clusters fill up and are used profitably. If utilisation lags, cost burden becomes heavy. So model utilisation scenarios: pessimistic, base, optimistic. - Regulatory / Geopolitical Discount Factor
Since Nebius came from Yandex, and sanctions impacted that, one needs to include risk discount for any potential regulatory actions, supply chain / chip availability risks etc. Could be non-linear risk. - Margin of Safety
Because of uncertainty, you’d want to buy well below what you believe is intrinsic value. So use lowest reasonable fair value you calculate, and aim for price < that. - Competitive Moat Stress Test
Ask: what prevents Microsoft, Amazon, or Google from doing the same thing (or better)? If they decide to push heavily on AI infrastructure, will Nebius be marginalized? Evaluate Nebius’s unique advantages.
Horizons 
Leave a comment