2 Artificial Intelligence Stocks That Can Have Their Nvidia Moment in 2026

AI-focused names CoreWeave and AMD are highlighted as potential 'next Nvidia' candidates based on strong top-line growth and product road maps. CoreWeave (CRWV) — trading only since March and volatile after a large run-up and ~60% drop from its high — reported Q3 2025 revenue near $1.4 billion, up 134% year-over-year, a Q3 net loss of $110 million (vs. $389 million a year ago), a P/S near 7 and $1.9 billion in liquidity with management anticipating ~136% revenue growth in 2026. AMD posted Q3 2025 revenue above $9.2 billion (up 36% YoY), with data center representing ~47% of revenue, net income over $1.2 billion (up 61% YoY), guidance for a 35% revenue CAGR over the next 3–5 years and an MI450 accelerator due in H2 next year; the shares are up roughly 60% over the last year with a current P/E of 106 and forward P/E of ~54.

Analysis

Market structure: Specialized AI infrastructure (CRWV) and accelerator suppliers (NVDA, AMD) are primary beneficiaries as customers pay premium for performance and low-latency stacks; legacy cloud incumbents face higher capex and potential margin pressure as workloads migrate to optimized providers. Fast revenue signals — CoreWeave +134% YoY Q3 2025 and AMD +36% Q3 2025 — imply demand >> available optimized capacity for at least 12–24 months, supporting pricing power for chips and colo capacity.

Risk assessment: Key tail risks are geopolitical/export controls on leading-edge nodes (TSMC/ASML restrictions) and a 20–40% cyclical pullback in AI spend if macro slows; CoreWeave’s $1.9bn liquidity reduces immediate bankruptcy risk but quarterly losses ($110m in Q3) create path-to-profitability dependency within 4–8 quarters. Near-term (days/weeks) momentum and H2 2026 product cadence (AMD MI450) drive flows; long-term (2026–2033) outcome tied to sustained 30%+ AI CAGR assumptions.

Trade implications: Tactical overweight semiconductors and AI-tailored cloud infra, underweight legacy-margin-exposed SaaS if margins compress; implement size-controlled positions (see decisions) and use calendar/vertical spreads around product launches to cap theta. Monitor GPU supply updates, large hyperscaler contracts, and Qs with ±15% share-price reaction windows as execution catalysts.

Contrarian angles: Consensus underestimates concentration risk — a few hyperscalers could internalize accelerators and undercut third-party demand, rapidly compressing CRWV-like valuations. Valuation dispersion is wide: AMD’s forward P/E ~54 prices growth but leaves vulnerability if MI450 underdelivers; historical parallels to past hardware cycles (2017–2019 GPU cycles) warn of 40–60% drawdowns amid supply normalization.

Key Decisions for Investors

• Establish a 2–3% long position in AMD (ticker AMD) within 2 weeks, target +40–60% upside over 9–12 months driven by MI450 adoption (H2 2026) and data‑center CAGR; place a hard stop-loss at -20% and trim half at +50% to lock gains.
• Allocate 1–1.5% to CoreWeave (CRWV) via staggered buy or 12–18 month call spreads (buy 1.5x ATM, sell 25% OTM) on pullbacks to P/S ≈5–6 or a 25–30% price decline; cap downside (stop or defined spread loss) at -30% given volatile drawdowns and $1.9bn liquidity runway.
• Buy asymmetric NVDA (NVDA) exposure with a 9–12 month call spread sized 1.5–2% of portfolio (buy ATM, sell 20–30% OTM) to participate in upside while limiting premium decay; increase if supply constraints persist after next two earnings prints.
• Implement a pair trade: Long AMD 2% / Short MSFT 1.5% over 3–9 months to express relative data-center share gain risk; unwind or rebalance if spread moves >15% adverse or after MI450 launch results are disclosed.