NBIS vs. GOOGL: Which AI-Infrastructure Play is the Better Buy Now?

Nebius is rapidly scaling AI data-center capacity, targeting 2.5 GW of contracted power by 2026 and announcing major hyperscaler deals — a $3 billion five-year contract with Meta and a $17.4–$19.4 billion agreement with Microsoft — while targeting $7–$9 billion ARR for 2026 and $900M–$1.1B by year-end 2025; however, NBIS has tightened 2025 revenue guidance to $500–$550M, raised capex guidance from ~$2B to ~$5B, and saw SG&A jump 87% YoY, leaving full-year adjusted EBITDA negative. Alphabet’s Google Cloud grew 33.5% YoY in Q3 2025, is accelerating AI initiatives with Gemini and custom GPUs, but is increasing capex guidance to $91–$93B for 2025 and faces stiff competition from AWS/Azure. Market reaction has been bifurcated (NBIS down ~24.3% and GOOGL up ~16.6% over the past month) and valuation/analyst metrics favor Alphabet (Zacks ranks: NBIS #4 Sell; GOOGL #3 Hold; NBIS price/sales ≈65.15 vs GOOGL ≈10.13).

Analysis

Market structure: Hyperscalers (GOOGL, MSFT, META) and GPU suppliers (NVDA) are the primary beneficiaries as AI shifts from models to raw compute and power — evidenced by NBIS aiming for 2.5 GW contracted by 2026 and GOOGL Cloud +33.5% YoY. Colocation pure-plays and small data‑center builders without hyperscaler commitments are at risk as pricing power consolidates with vertically integrated providers and GPU vendors; power and copper markets will tighten regionally as data‑center builds accelerate.

Risk assessment: Key tail risks are regulatory export controls on GPUs and antitrust action, grid permitting delays, and large contract cancellations or re-negotiations (NBIS’s $3B Meta and ~$17.4–$19.4B Microsoft deals may have timing/recognition clauses). Near term (days–weeks) watch earnings and capex guidance; medium term (3–12 months) execution on NBIS capacity; long term (1–3 years) demand elasticity, GPU architecture shifts and energy constraints could flip winners into losers.

Trade implications: Favor large-cap hyperscalers and GPU suppliers; avoid/short speculative infra names with stretched P/S (NBIS P/S ~65) until contracted ARR converts to cash. Tactically, use pairs (long GOOGL, short NBIS) and options to limit downside: buy GOOGL 12‑18 month calls or LEAPS and purchase NBIS 3–6 month put spreads; overweight NVDA exposure to play ongoing GPU scarcity and price-inelasticity.

Contrarian view: Market may be underpricing execution risk in NBIS despite headline contracts — those dollars depend on power, supply chains, and multi-year capex (NBIS capex guide jump to ~$5B in 2025). Conversely GOOGL’s heavy capex ($91–93B in 2025) is being discounted as growth; history (early AWS) shows winners often outspend peers and take short-term margin pain for durable share — credit that possibility while sizing risk.

Key Decisions for Investors

• Establish a 1.5% portfolio long in GOOGL (ticker GOOGL) over 12–18 months targeting +15–25% upside if Google Cloud sustains >25% ARR growth; set a protective stop-loss at -12% or trail 10% after a +8% move.
• Initiate a pair trade: short NBIS (0.5–1.0% notional) while going long GOOGL of equal notional; if options available, buy NBIS 3–6 month put spread (sell -5% lower strike) to cap premium and buy GOOGL 12 month LEAPS call 10% OTM for asymmetric upside.
• Buy a directional NVDA position (0.5–1.0% of portfolio) via a 3–6 month call spread to express continued GPU tightness; target +30–50% if Blackwell adoption accelerates, exit if NVDA misses next supply/guide or implied vol pops >+40% intraday.
• Reduce exposure by 30–50% to small-cap pure-play data-center/colo builders (names lacking hyperscaler take-or-pay contracts) and reallocate into utilities/grid equipment suppliers and selected hyperscalers over the next 60 days to hedge energy/permitting risk.