Nebius annual revenue

Calculation of starting probabilities

This market uses a three-level revenue model for Nebius's non-cumulative annual revenue over FY2026-FY2035. The model combines: (1) Nebius-specific execution signals, (2) a comparable cloud-platform reference class, and (3) a general firm-growth base rate from the academic literature so the forecast does not naively extrapolate one explosive year forever.

Resolved anchor from the last completed financial year: Nebius reported $529.8m of FY2025 revenue versus a recast FY2024 continuing-operations revenue of $91.5m, which is +479% year over year. That value is included here as a calibration aid; the market itself stays in revenue units only.

Key charts

Why these reference classes

• Nebius-specific: FY2025 revenue, year-end ARR of about $1.25bn, FY2026 end-year ARR guidance of $7bn-$9bn, the Microsoft agreement announced in December 2025, and disclosed power-capacity targets (800MW-1GW active in 2026; more than 3GW contracted).
• Broad cloud/platform cohort: Cloudflare, DigitalOcean, Datadog, and Snowflake. Their observed mean annual log-growth in the sample used here is about 32%, which prevents the model from treating Nebius as a generic mature company too early.
• AI-infrastructure uplift: CoreWeave is used as the direct AI-cloud analogue, but only after strong shrinkage. Its raw growth history is much hotter (mean about 587%), so the model only takes a partial, decaying uplift from that reference class to reduce single-comparable bias.
• Guidance-error outside view: the FY2026 anchor is built from management's end-2026 ARR guidance converted to a FY2026 revenue center (about $4,441m before pooling). The model then gives this guidance layer about 87% average weight in FY2026 because near-term guidance has higher predictive power than unconstrained long-horizon structural simulation.
• Companies in general: the academic literature says firm growth shocks are heavy-tailed, high growth is rarely persistent forever, and volatility falls with size. That is why the simulation uses fat-tailed shocks and size-scaled volatility instead of a smooth deterministic curve.

Model structure

1. Global layer: long-run revenue growth mean-reverts toward a modest positive floor, with shock volatility shrinking roughly with firm size.
2. Sector layer: Nebius inherits a broad-cloud growth premium estimated from comparable public firms, plus a shrunk AI-infrastructure uplift that decays as the market matures.
3. Company and guidance layer: FY2026 first forms a structural path from the global+sector model, then precision-pools it with a guidance-derived revenue center using inverse-variance weights. That keeps the outside view, but prevents excessive mass far below near-term guidance.
4. Simulation: 80,000 Monte Carlo paths produce yearly revenue samples, and each year's market ladder is the survival curve P(revenue >= threshold) at 50 friendly-rounded thresholds.

Forecast summary

Source notes

• Nebius FY2025 results: FY2025 revenue of $529.8m and recast FY2024 revenue of $91.5m.
• Nebius Q3 2025 shareholder letter: year-end ARR of about $1.25bn, FY2026 end-year ARR guidance of $7bn-$9bn, and power-capacity targets.
• Nebius-Microsoft agreement: direct evidence that the company has unusually strong near-term demand support.
• Stanley et al. (1996) and Bottazzi & Secchi (2006): motivation for fat-tailed growth shocks and size-sensitive volatility.
• World Bank high-growth firms review: motivation for mean reversion and caution about assuming repeated hypergrowth episodes.

Data and reproducibility assets

• Download the model summary JSON
• Download the forecast chart
• Download the reference-class chart

Download the Python script used for this upload