Nebius (NBIS), the rebranded Yandex cloud unit, has posted explosive top-line growth — revenue +462% in 2024 and +437% y/y in the first nine months of 2025 — and trades around $97 with a $23bn market cap (under 7x 2025 sales), but it runs only one owned data center, relies heavily on Microsoft and Meta, must scale capex to expand and is expected to incur steep losses in 2026–27. The piece counsels a more conservative exposure to AI infrastructure via DigitalOcean (DOCN), a $5.4bn company with a 2021–24 revenue CAGR of 22%, GAAP profitability since 2023 and analysts forecasting ~15% revenue growth in 2025 and 19% revenue / 15% adj. EPS CAGRs through 2027, trading at ~5x sales and ~29x forward adjusted earnings.
Market structure: Nebius (NBIS, market cap ~$23B) is a high-growth specialist capturing hyperscaler GPU/cloud demand (462% rev growth in 2024, 437% YTD 2025) but operates one owned DC and multiple colocations — a model that amplifies scale-up capex and concentration risk. DigitalOcean (DOCN, $5.4B) is a defensive small-to-mid-business cloud play (22% CAGR 2021–24, GAAP profitable since 2023) positioned to capture SMB AI workloads; hyperscalers (MSFT, AMZN, META) remain both customers and competitive pricing anchors. Pricing power will bifurcate: hyperscalers hold leverage to compress rents; niche providers with committed long-term contracts or unique geography/energy costs can command premiums. Risk assessment: Tail risks include renewed geopolitical sanctions or export controls on AI hardware, a GPU supply squeeze, or NBIS failing to finance rapid DC buildouts (analysts expect steep losses in 2026–27). Short-term (days–months) volatility will hinge on contract disclosures and quarterly capex guidance; medium-term (6–18 months) hinge on data-center commissioning and customer diversification; long-term (2–3 years) on sustainable gross margins and low customer concentration (<30% for any single client). Hidden dependencies: NBIS’s leased-colocation footprint exposes it to counterparties’ power/contract risk and FX/electricity swings in Europe. Trade implications: Tactical posture — underweight NBIS vs overweight DOCN. Establish a modest long DOCN (2–3% portfolio) to capture steady AI SMB adoption and add a hedged short NBIS (1–2%) or buy downside protection; use a 3–6 month put spread on NBIS (buy 1 6‑month 95 put, sell 1 6‑month 65 put) sized to 1% notional to limit cost. Consider pair-trade: long DOCN equal-notional short NBIS to isolate execution/capex risk; rotate proceeds into NVDA or MSFT for platform exposure if NVDA remains >20% of AI hardware upside. Contrarian angles: The market may be over-counting NBIS’s rebrand narrative and underweighting capex dilution and client concentration; conversely it may underprice the value of European low-carbon/sovereign infrastructure (Iceland/Finland) should energy-cost differentials widen. Historical parallel: early cloud specialists that grew fast but lost pricing discipline often required equity raises and saw >50% dilution; if NBIS’s capex/revenue exceeds ~40–50% in 2026, treat equity as highly dilutive. Watch triggers: any single-client revenue >30%, quarterly free-cash-flow turning negative by >20% of market cap, or announced multi‑year hyperscaler price concessions.
AI-powered research, real-time alerts, and portfolio analytics for institutional investors.
Request a DemoOverall Sentiment
mixed
Sentiment Score
-0.10
Ticker Sentiment
