2 Genius Stocks Nvidia Owns That You Should Buy for 2026

Nvidia disclosed 13F holdings including ~7.7 million shares (≈2.8%) of Applied Digital and ~1.2 million shares (≈0.5%) of Nebius, reflecting exposure to the AI-driven data-center buildout. Applied Digital is up ~238% YTD 2025 after long-term AI lease announcements, while Nebius projects ARR of $900M–$1.1B by end-2025 and $7B–$9B by 2026; both companies remain unprofitable as they scale. The stakes are small relative to Nvidia but signal strategic interest in fast-growing infrastructure providers, presenting higher-risk, higher-reward opportunities versus Nvidia’s profitable, lower-risk profile.

Analysis

Market structure: Winners are GPU manufacturer Nvidia (NVDA) and AI-focused data-center operators (APLD, NBIS, CRWV/CoreWeave) that can secure long-term GPU capacity and 10–15+ year leases; losers are commodity colocation providers and legacy IT outsourcers unable to capture AI premium. The near-term pricing power sits with GPU suppliers and integrated operators—spot GPU rents and multi-year capacity contracts can sustain 30–50%+ revenue premiums versus pre-AI levels for 12–24 months. Cross-asset effects include tighter credit spreads for well-contracted operators, upward pressure on industrial commodities (copper, power), and higher implied vol on NVDA options; duration-sensitive tech credits could reprice if capex-led cash burn accelerates.

Risk assessment: Tail risks include export controls on high-end GPUs (BIS action within 3–6 months could cut revenue 20–60% for exposed names), severe build execution failures causing equity dilution (equity raises >20%), or an AI demand slowdown that reduces ARR growth to single digits by 2026. Time horizons: immediate (days) volatile on 13F/news flow, short-term (weeks–months) repricing around earnings and capacity updates, long-term (2–3 years) depends on profitability conversion and power/real-estate constraints. Hidden dependencies: GPU wafer/allocations from NVDA, long-term power contracts, and counterparty credit of anchor tenants.

Trade implications: Core trade is conviction long NVDA (2–4% portfolio) for 6–12 months with buy-and-trim discipline; speculative stakes in APLD and NBIS (0.5–1% each) with strict 35–40% stop-loss or protective puts. Pair: long NBIS or APLD vs short legacy colo (e.g., EQIX) to express AI-premium vs commodity colocation; options: use 3–9 month call spreads on NVDA to cap cost and buy LEAP puts as tail-hedge. Rotate 2–4% from broad tech into infrastructure/integrated operators over 4–12 weeks.

Key Decisions for Investors

• Establish a 2–4% portfolio long in NVDA (ticker NVDA) as core AI exposure with a 6–12 month horizon; trim 25–50% on a +40% move and cut to reduce to 1% if the position falls -20% from entry.
• Initiate small, speculative positions in APLD and NBIS (0.5–1% each) funded from reduced broad-tech beta; protect each with either a 35–40% hard stop or buy 6–9 month puts (cost ≤0.5% portfolio each) to limit downside from execution or GPU-export shocks.
• Implement a relative-value pair: long NBIS or APLD (equal weight 1% total) vs short legacy colo REIT EQIX (0.5–1%), expressing AI-contract premium; rebalance if spread compresses >15% or after next quarterly ARR prints (within 60–90 days).
• Use options to express asymmetric risk: buy 3–9 month NVDA call debit spreads sized to 0.5–1% portfolio to capture upside while selling 1–3 month covered calls on existing NVDA holdings to monetize theta; purchase 12–18 month LEAP puts (tail hedge) if portfolio NVDA exposure >3%.
• Monitor catalysts over next 30–90 days and act: BIS/US export policy updates, NVDA earnings and guidance, quarterly ARR and contracted MW disclosures from NBIS/APLD; if export controls tighten materially, reduce GPU-exposed long bets by 50% within 7 trading days.