Nebius, Like AI, Is Evolving

Nebius Group says its AI infrastructure business is now scaling profitably, with adjusted EBITDA swinging sharply positive. The company is expanding beyond GPUs into inference, agentic AI, and enterprise AI deployment, while targeting $7B-$9B in ARR by 2026. The article frames a $300 share price for NBIS as increasingly realistic, signaling a much stronger long-term growth outlook.

Analysis

The key signal here is not just improving profitability, but that the business is crossing from a capital-intensive GPU rental model into a software-like consumption model. If NBIS can monetize inference and agentic workloads, gross margin should expand as utilization rises and customer workloads become stickier; that usually compresses payback periods on deployed capex and supports a rerating multiple, not just higher revenue. The market is likely underestimating how quickly enterprise AI deployment can become the dominant driver once proof points emerge, because the sales cycle is longer but the resulting contracts are far more durable.

The main winners are likely the next layer of AI infrastructure suppliers: networking, power, cooling, and specialized components that benefit as inference workloads scale beyond training. The losers are the “commodity GPU rental” competitors whose economics stay exposed to price competition and capacity glut; once customers start buying integrated deployment rather than raw compute, standalone GPU brokers get disintermediated. Second-order, the more NBIS proves margins can scale, the more it pressures hyperscalers and neocloud peers to show their own utilization and pricing power, which can widen dispersion across the AI infrastructure group.

The risk is that the market extrapolates the EBITDA inflection too aggressively into 2026 ARR without enough evidence on customer concentration, renewal durability, and capex intensity. A reversal would likely come from a modest slowdown in enterprise AI adoption, inference pricing pressure, or any sign that revenue growth is being bought with unsustainably high deployment spend. Near term, the stock can keep working on momentum over days to weeks, but the fundamental test is over the next 2-4 quarters: whether new contracts convert into repeat usage and not just headline ARR.

Key Decisions for Investors

• Go long NBIS on pullbacks over the next 1-2 weeks; use the recent EBITDA inflection as a catalyst, but size for elevated multiple risk. Target a 2-3x upside on a successful rerating, with a hard stop if enterprise bookings or margin expansion decelerate in the next quarterly update.
• Pair trade: long NBIS / short a weaker GPU-rental or capital-heavy AI infra peer over 1-3 months. The trade benefits if the market starts rewarding mix shift and operating leverage rather than pure revenue growth.
• Buy medium-dated call spreads on NBIS to express upside to the $300 narrative while capping premium burn. Best structure is 6-12 months out, where catalyst path is clearer and implied volatility should still be manageable.
• Monitor networking/power infrastructure names for second-order upside over the next quarter; if NBIS continues to prove demand quality, the trade is to own the picks-and-shovels rather than chase the headline winner after the rerating has already occurred.
• Do not short into strength unless there is evidence of capex overspend or customer churn; the better contrarian stance is to wait for a failed post-earnings gap or guidance reset before considering a fade.