Nebius Group (NBIS) The Full-Stack AI Cloud Hiding in Plain Sight

Investment Thesis

Nebius Group N.V. is one of the most compelling asymmetric AI infrastructure opportunities in public markets today. The company — a vertically integrated, full-stack AI cloud built on GPU-as-a-Service (GPUaaS) — delivered $529.8M in FY2025 revenue (+479% YoY) and exited the year with $1.25B in annualized recurring revenue (ARR), beating its own guidance. The 2026 revenue guide of $3.0–3.4B implies a 5–6x scale-up in a single fiscal year, backed by signed hyperscaler contracts, a $2B NVIDIA strategic investment, and a 3GW pipeline of owned and controlled compute capacity.

Demand is structurally exceeding supply. Year-long contracts doubled quarter-over-quarter. Average selling prices expanded 50% sequentially. Gross margins reached 70% in Q4 2025. CEO Arkady Volozh (co-founder of Yandex) and CRO Marc Boroditsky confirmed at the March 2026 Morgan Stanley TMT Conference that the bottleneck is not demand — it is the physical rate at which new data centers come online. That is a rare and powerful position for any infrastructure company.

Three drivers anchor the BUY: (1) ARR ramp toward $7–9B exit 2026, which at 6–8x ARR implies a $42–72B market cap vs. $28.4B today; (2) NVIDIA's $2B equity investment as a structural validator and preferential supply agreement; (3) Independence, Missouri data center — 800MW expandable to 1.2GW — coming online in phases across 2025–2026, providing the capacity to execute on the contracted backlog. We initiate with a BUY and 12-month price target of $155, representing 37% upside.

Sector Snapshot — Key Metrics

Quarterly Revenue Trend

ARR Progression

Nebius exited Q4 2025 with $1.25B ARR, having grown from roughly $250M a year prior — a 5x year-over-year acceleration. The $7–9B ARR exit target for 2026 reflects signed commitments from hyperscale customers and a rapid cadence of new data center capacity coming online across the U.S. and Europe. This trajectory, if achieved, would put Nebius among the fastest-growing infrastructure companies ever.

Business Model & Competitive Positioning

Nebius operates as a full-stack AI cloud, providing GPU compute, AI tooling, managed services, and proprietary ML infrastructure across its owned data centers. Unlike colocation providers, Nebius controls the entire stack — hardware procurement (Tier-1 NVIDIA supplier), data center design, software layer (Nebius AI Studio, vector databases, model inference APIs), and customer support.

Core revenue segments include GPU cloud (billed per-GPU-hour on reserved and on-demand basis), managed AI platform services, and long-term capacity agreements with hyperscale customers. The hyperscaler agreements — including a reported ~$19.4B capacity commitment from Microsoft and ~$3B from Meta — de-risk the CapEx program by pre-committing revenue before facilities are built.

Competitive differentiation rests on three pillars: (1) GPU supply access via NVIDIA's $2B equity stake, providing preferential allocation of H100/H200/B200 clusters; (2) owned infrastructure, enabling significantly higher margins than leased alternatives; (3) purpose-built AI tooling that increases customer switching costs and expands wallet share per customer.

Financial Snapshot

Competitive Benchmark

Nebius leads on AI/GPU focus and cost efficiency relative to public-cloud giants, but trails on revenue scale and customer breadth. CoreWeave is the most directly comparable peer (pure-play GPU cloud). AWS and Azure dominate on scale and diversification, but their GPUaaS offerings are bundled within broader cloud contracts, limiting comparability. Lambda Labs remains early-stage in revenue scale. Nebius's overall score of 58 is expected to improve materially as the 2026 ARR ramp converts contracted backlog into recognized revenue.

Key Catalysts (12-Month Horizon)

• Independence, MO data center Phase 1 full commissioning (mid-2026) — 300–400MW online
• Q1 2026 earnings: first large-scale revenue recognition from hyperscaler take-or-pay agreements
• Additional NVIDIA GPU allocation announcements under the $2B strategic investment umbrella
• ARR tracking updates — any print exceeding $2B exit on a quarterly basis would re-rate the stock significantly
• Potential European data center announcements adding to the 3GW pipeline
• New long-term customer wins beyond Microsoft/Meta — diversification of the revenue base

Key Risks

• CapEx execution: $16–20B of planned 2026 spend is enormous; construction delays or cost overruns could compress margins
• Customer concentration: Microsoft and Meta represent the majority of contracted backlog; churn would be highly impactful
• GPU supply: Despite NVIDIA partnership, AI chip demand is industry-wide; supply allocation could face competition
• Debt load: $4.89B in total debt against $3.7B cash requires disciplined capital allocation and successful ARR ramp
• Geopolitical/regulatory risk: Nebius is a Yandex spin-off incorporated in the Netherlands; ongoing EU/US regulatory attention
• Macro slowdown: Hyperscaler AI CapEx could be dialed back if cloud spending growth decelerates across the industry

Valuation & Price Target

Our $155 price target is based on a blended ARR multiple and forward revenue analysis. Using a 5x multiple on our 2026E ARR midpoint of $8B yields an $40B enterprise value; adjusting for net debt (~$1.2B net) and applying a modest growth premium given the NVIDIA partnership, we arrive at $155 on a fully diluted basis (~252M shares). A 6x ARR multiple (justified by 70% gross margins and demand-supply imbalance) implies $210+.

The street high of $220 reflects a bull case where Nebius executes on the $10B+ ARR scenario, margins expand above 75%, and NVIDIA deepens the strategic relationship through further equity ownership. The bear case of $65 assumes CapEx delays push ARR to $3.5B, margins compress on underutilization, and the stock de-rates to a 1.5x forward revenue multiple.

Scenario Analysis

The NVIDIA Signal

NVIDIA's $2B strategic investment in Nebius (announced March 2026) is more than a financial transaction — it is a supply chain endorsement. By taking an equity stake, NVIDIA aligned its own returns with Nebius's success, which means preferential GPU allocation at scale is not a future hope but a current structural advantage. For a company whose bottleneck is compute supply rather than customer demand, this is the single most important de-risking event in the company's history.

Historically, NVIDIA has made equity investments in companies it views as critical distribution partners for its GPU ecosystem. The playbook: anchor the neocloud, ensure GPU clusters are deployed at maximum velocity, accelerate software adoption (CUDA, NIM, AI Enterprise), and benefit as the customer base grows. Nebius becomes a flagship showcase for NVIDIA's full-stack AI infrastructure — a role that brings both supply priority and marketing leverage for enterprise customer acquisition.

Frequently Asked Questions

Why is Nebius different from other neoclouds?

Nebius owns its data centers (not colocation), controls its hardware procurement through the NVIDIA relationship, and has built a proprietary AI software layer. This vertical integration enables 70% gross margins — dramatically higher than asset-light competitors — and creates compounding switching costs as customers integrate Nebius AI Studio, vector DBs, and inference APIs.

Isn't the $4.89B debt load concerning?

The debt is largely matched by contracted ARR and hyperscaler take-or-pay agreements. With $3.7B in cash and $3.0–3.4B in guided 2026 revenue, Nebius has sufficient liquidity to service obligations while executing the CapEx program. The risk is execution: delays compress liquidity. We model a base case where debt-to-EBITDA peaks in H1 2026 and normalizes rapidly as capacity comes online.

Can Nebius really achieve $7–9B ARR by year-end 2026?

It is aggressive but achievable. The key input is new data center capacity: if the Independence, MO campus delivers 600MW+ by H2 2026 on schedule, and European expansions add another 200–300MW, $7–9B ARR is within reach given current ASPs and contract structure. CEO Volozh confirmed at the Morgan Stanley conference that demand is 'far exceeding supply' — the constraint is buildout speed, not revenue appetite.

Conclusion — Initiate BUY, PT $155

Nebius Group is a rare opportunity to buy a hyper-growth AI infrastructure platform at an early-stage revenue multiple before the largest single-year revenue ramp in its history. The combination of NVIDIA backing, owned infrastructure, 70% gross margins, hyperscaler-anchored ARR contracts, and a CEO who built one of Europe's most valuable internet companies creates a risk/reward profile that is compelling at current prices.

We initiate coverage with a BUY rating and a 12-month price target of $155 (+37% from $112.95). Upside scenarios toward $210–220 are achievable if CapEx execution is on-time and ARR tracking confirms the $7–9B exit path. Investors should size positions to account for execution risk in a capital-intensive buildout phase, but the structural tailwinds — demand for AI compute, supply scarcity, and the NVIDIA relationship — make NBIS one of the highest-conviction AI infrastructure names in the public market.

DISCLAIMER: This report is produced by Edgen Research for informational purposes only and does not constitute investment advice. All financial data sourced from company filings, 360 Research (March 2026), and the Morgan Stanley TMT Conference transcript (March 4, 2026). Price data as of March 13, 2026. Past performance is not indicative of future results. Edgen Research and its affiliates may hold positions in securities mentioned.

Edgen Research · Published: March 17, 2026 · Analyst: Edgen AI Platform